docs(v1.1.5): reviewer audit report — APPROVE verdict

Audit of feat/v1.1.5-files-pubsub against the v1.1.5 dispatch prompt.
All gates green on HEAD; 491 tests pass (+64 new), 139 ignored.

Atomic write protocol audited line-by-line: single-pass SHA-256,
temp→fsync→rename→fsync-dir→DB sequence as specified, unique pid+
counter temp suffix, path-traversal defense at SDK boundary and repo.
Pub/sub fan-out is correctly transactional (single tx begin+commit;
one outbox row per matching subscriber; trigger_depth saturating-
bumped). Topic pattern matcher rejects every shape the brief called
out (*.created, **, a.*.b, user.*x, *user, empty).

Three flagged open questions resolved: orphan-sweep deferred (matches
planning decision), test count 63 vs 70 (defensible — gap is the
dispatcher e2e test, which is already covered for kv/docs/cron via
the shared dispatcher path), empty-blob = missing-data (defensible
interpretation, relaxable later).

First CI workflow added; schema_snapshot un-ignored with DATABASE_URL-
absent skip path.

.github/workflows/ci.yml

@@ -0,0 +1,72 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+env:
+  CARGO_TERM_COLOR: always
+  # Matches what docker-compose produces locally; the schema-snapshot
+  # guardrail and any other DB-backed tests run against this service.
+  DATABASE_URL: postgres://picloud:picloud@localhost:5432/picloud
+
+jobs:
+  rust:
+    name: Rust — fmt, clippy, test
+    runs-on: ubuntu-latest
+    services:
+      postgres:
+        image: postgres:15
+        env:
+          POSTGRES_USER: picloud
+          POSTGRES_PASSWORD: picloud
+          POSTGRES_DB: picloud
+        ports:
+          - 5432:5432
+        options: >-
+          --health-cmd "pg_isready -U picloud"
+          --health-interval 10s
+          --health-timeout 5s
+          --health-retries 5
+    steps:
+      - uses: actions/checkout@v4
+
+      # rust-toolchain.toml pins the channel; this action honors it.
+      - name: Install Rust toolchain
+        uses: dtolnay/rust-toolchain@stable
+        with:
+          components: rustfmt, clippy
+
+      - name: Cache cargo
+        uses: Swatinem/rust-cache@v2
+
+      - name: Format check
+        run: cargo fmt --all -- --check
+
+      - name: Clippy
+        run: cargo clippy --all-targets --all-features -- -D warnings
+
+      # Runs the whole workspace, including the schema-snapshot guardrail
+      # (it picks up DATABASE_URL from the env above and the postgres
+      # service; without a DB it would skip cleanly).
+      - name: Test
+        run: cargo test --workspace
+
+  dashboard:
+    name: Dashboard — check
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: dashboard
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-node@v4
+        with:
+          node-version: 20
+          cache: npm
+          cache-dependency-path: dashboard/package-lock.json
+      - name: Install deps
+        run: npm ci
+      - name: Svelte check
+        run: npm run check

CHANGELOG.md

@@ -1,5 +1,368 @@
 # PiCloud Changelog
 
+## v1.1.5 — Files & Pub/Sub (unreleased)
+
+Two stateful services + two trigger kinds. **`files::*`** is
+filesystem-backed blob storage (atomic writes, path-sharded layout,
+single-pass SHA-256 with checksum-verified reads); the metadata row
+lives in Postgres, the bytes on disk. **`pubsub::publish_durable`** is
+durable pub/sub through the universal outbox, fanning out one delivery
+row per matching subscriber **at publish time** inside a single
+transaction. Both ride the v1.1.1 trigger framework as the fifth and
+sixth concrete kinds via the established Layout-E extension pattern.
+
+### Added
+
+- **`files::collection(name).{create,head,get,update,delete,list}`** —
+  blob storage SDK. `create`/`update` take a Rhai `Blob`; `get` returns
+  a `Blob` (or `()` if missing); `head`/`list` return metadata maps
+  (`id, name, content_type, size, checksum, created_at, updated_at`).
+  `create`/`update`/`delete` throw on failure; `get`/`head` return `()`
+  for a missing file; `delete` returns a was-present bool. Missing
+  required field on `create` throws naming the field.
+- **Atomic writes** — temp file → fsync → rename → fsync parent dir →
+  DB row, so a crash never leaves a readable half-written file. SHA-256
+  is computed in a single pass during the write; `get` re-verifies it
+  and surfaces `FilesError::Corrupted` (logged with the path, never
+  auto-deleted) on a mismatch. Shard dirs are created `0o700`.
+- **`files:*` trigger kind** — `ctx.event.files` carries the metadata
+  only (never the bytes; a handler that wants them calls
+  `files::collection(c).get(id)`). `prev` is `()` on create, the prior
+  metadata on update, the deleted metadata on delete.
+- **`pubsub::publish_durable(topic, message)`** — durable publish.
+  Message is any JSON-serializable Rhai value; Blobs encode as base64
+  (at any nesting depth). No matching subscriber → the publish succeeds
+  silently with zero outbox rows.
+- **`pubsub:*` trigger kind** — topic patterns are exact, `<prefix>.*`,
+  or `*`; mid-pattern wildcards are rejected at trigger creation.
+  `ctx.event.pubsub` carries `topic`, `message`, `published_at`.
+- **`FilesService` + `PubsubService` traits** (`picloud-shared`) +
+  `FsFilesRepo`/`FilesServiceImpl` and `PostgresPubsubRepo`/
+  `PubsubServiceImpl` (manager-core). Wired into the `Services` bundle
+  as `files` and `pubsub`.
+- **Capabilities** `AppFilesRead`/`AppFilesWrite` → `script:read`/
+  `script:write`, `AppPubsubPublish` → `script:write`. No new `Scope`
+  variant — the seven-scope commitment holds. Script-as-gate: skipped
+  when the script runs unauthenticated.
+- **Admin files API** (`GET`/`DELETE /apps/{id}/files`) + dashboard
+  Files view per app; **Pub/Sub trigger form** on the Triggers tab.
+- **CI** — first `.github/workflows/ci.yml` (Postgres service, fmt +
+  clippy + `cargo test --workspace`); the schema-snapshot guardrail now
+  runs instead of being `#[ignore]`'d.
+
+### Changed
+
+- Workspace version: 1.1.4 → 1.1.5
+- Rhai SDK version: 1.5 → 1.6
+- Dashboard version: 0.10.0 → 0.11.0
+- `schema_snapshot` test: no longer `#[ignore]`'d — runs against
+  `DATABASE_URL` when set, skips cleanly when absent.
+
+### Migrations
+
+- 0018_files.sql — `files` metadata table (bytes live on disk).
+- 0019_files_triggers.sql — widen kind/source_kind CHECKs + add
+  `files_trigger_details`.
+- 0020_pubsub_triggers.sql — widen kind/source_kind CHECKs + add
+  `pubsub_trigger_details` + partial index.
+
+### New environment variables
+
+- `PICLOUD_FILES_ROOT` (default `./data`)
+- `PICLOUD_FILES_MAX_FILE_SIZE_BYTES` (default 100 MB)
+
+## v1.1.4 — Outbound HTTP & Cron triggers (unreleased)
+
+Two surfaces. **`http::*`** lets Rhai scripts make outbound HTTP
+requests (Slack webhooks, Stripe, third-party REST) fronted by an SSRF
+deny-list applied to the *resolved IP* (DNS-rebinding defense), with
+scheme/port restrictions, request/response body caps, and a layered
+timeout. **Cron triggers** add the fourth concrete kind on the v1.1.1
+trigger framework: a scheduler task enqueues due triggers into the same
+universal outbox the dispatcher already drains.
+
+### Added
+
+- **`http::{get,post,put,patch,delete,head,post_form,request}`** — outbound
+  HTTP SDK. Body and options are separate positional args
+  (`verb(url, body, opts)`); `opts` is
+  `{headers, timeout_ms, follow_redirects, max_redirects}` (unknown keys
+  throw). Body dispatch by type: Map/Array → JSON, String → text/plain,
+  `()` → none. Response is `#{ status, headers, body, body_raw }` with
+  `body` auto-parsed when the response is `application/json`. Non-2xx
+  does NOT throw (fetch-style); network/timeout/SSRF/size errors throw
+  with an `"http: …"` prefix.
+- **SSRF deny-list** — applied to the resolved IP via a custom reqwest
+  `dns_resolver` (so it covers every redirect hop and defeats DNS
+  rebinding), plus a literal-IP check at URL-parse time. Blocks
+  loopback, RFC1918 private, link-local (incl. `169.254.169.254`),
+  carrier-grade NAT, multicast, reserved, IPv6 ULA/link-local/loopback,
+  and IPv4-mapped IPv6 (re-checked against the embedded v4 address).
+  The script-visible error carries a CIDR-category reason, never the IP.
+  `PICLOUD_HTTP_ALLOW_PRIVATE=true` disables it (dev-only; logs a startup
+  warning).
+- **`HttpService` trait** (`picloud-shared`) + `HttpServiceImpl`
+  (manager-core, reqwest-backed). Wired into the `Services` bundle as
+  `http: Arc<dyn HttpService>`.
+- **`Capability::AppHttpRequest(AppId)`** — maps to the existing
+  `script:write` scope (any outbound request can exfiltrate data, so the
+  conservative write mapping is used). No new `Scope` variant — the
+  seven-scope commitment holds. Script-as-gate: skipped when the script
+  runs unauthenticated.
+- **Cron triggers** — `POST /api/v1/admin/apps/{id}/triggers/cron`
+  (`script_id`, `schedule`, `timezone`, optional retry overrides).
+  6-field cron expressions (with seconds) validated by the `cron` crate;
+  IANA timezones validated by `chrono-tz`. A scheduler task
+  (`spawn_cron_scheduler`, poll cadence `PICLOUD_CRON_TICK_INTERVAL_MS`,
+  default 30s) enqueues due triggers into the outbox; the existing
+  dispatcher delivers them. Catch-up policy: a trigger that missed N
+  windows fires exactly **once** on the next tick, not N times.
+- **`ctx.event.cron`** — `{ schedule, timezone, scheduled_at, fired_at }`
+  for cron-trigger handlers (`ctx.event.source == "cron"`,
+  `ctx.event.op == "tick"`).
+- **Dashboard Triggers tab** — admin-gated cron trigger create form
+  (target endpoint script, schedule, timezone dropdown) + triggers list
+  showing schedule / timezone / last-fired.
+
+### Changed
+
+- **Workspace version**: `1.1.3` → `1.1.4`.
+- **Rhai SDK version**: `1.4` → `1.5` (additive — `http::*` SDK +
+  `ctx.event.cron`). The `Services` bundle constructor becomes
+  `Services::new(kv, docs, dead_letters, events, modules, http)`.
+- **Dashboard version**: `0.9.0` → `0.10.0`.
+- **`SdkCallCx`** — gains a `script_id` field (audit attribution + the
+  default outbound `User-Agent`, `picloud/<version> (script:<id>)`).
+- **Rhai pin tightened** — workspace dep `rhai = "1.19"` → `rhai = "=1.24"`
+  so future bumps of the non-semver-stable `internals` surface are
+  deliberate.
+- **Module backend errors redacted** — `PicloudModuleResolver` now
+  surfaces a stable generic (`"module backend unavailable; check server
+  logs"`) to scripts and logs the original at error level, instead of
+  leaking the backend error verbatim (see v1.1.3 follow-up).
+
+### Migrations
+
+- `0017_cron_triggers.sql` — widens `triggers.kind` and
+  `outbox.source_kind` CHECK constraints to include `'cron'`; adds
+  `cron_trigger_details (trigger_id, schedule, timezone, last_fired_at)`
+  with a `last_fired_at` index. Additive — applies cleanly on a fresh DB
+  and on top of the v1.1.3 schema.
+
+### New environment variables
+
+- `PICLOUD_HTTP_ALLOW_PRIVATE` (default false; dev-only) — disable the
+  SSRF deny-list.
+- `PICLOUD_HTTP_MAX_REQUEST_BODY_BYTES` / `PICLOUD_HTTP_MAX_RESPONSE_BODY_BYTES`
+  (default 10 MB each).
+- `PICLOUD_CRON_TICK_INTERVAL_MS` (default 30000) — cron scheduler poll
+  cadence (floored at 1s).
+
+## v1.1.3 — Modules (unreleased)
+
+Real per-app Rhai module system. Scripts can `import "<name>" as
+<alias>;` other scripts in the same app as reusable libraries. The
+v1.0 placeholder `DummyModuleResolver` is replaced by a per-call
+`PicloudModuleResolver` that loads `kind = 'module'` scripts via a
+new `ModuleSource` trait, compiles them into Rhai modules, caches
+the compiled output, and enforces cross-app isolation, circular-
+import detection, and an import-depth limit. Two LRU AST caches
+(top-level script + per-module compiled module) eliminate the
+per-invocation compile cost; both invalidate on `updated_at` change.
+
+### Added
+
+- **`scripts.kind` column** — `'endpoint' | 'module'`, default
+  `'endpoint'`. Endpoints handle HTTP routes / trigger events;
+  modules are libraries imported by other scripts. The dashboard
+  scripts list + script detail page surface the distinction as a
+  colored badge.
+- **`script_imports` dep-graph table** — populated at script save-
+  time from the literal-path `import "<name>"` declarations in the
+  source. FK-CASCADE on both columns. No admin surface in v1.1.3
+  (drives a v1.2+ "Used by" dashboard panel and v1.3+ cluster-mode
+  eager invalidation).
+- **`ModuleSource` trait** — `lookup(&SdkCallCx, name)`. Postgres
+  impl `PostgresModuleSource` in manager-core. `app_id` derived from
+  `cx.app_id` (cross-app isolation boundary, mirrors KV / docs).
+- **`PicloudModuleResolver`** — implements `rhai::ModuleResolver`.
+  Per-call instance owns `Arc<SdkCallCx>`, the in-progress imports
+  stack, the depth counter. Bridges sync `resolve()` to async
+  `lookup()` via `Handle::block_on` (safe under the executor's
+  `spawn_blocking` wrap). Replaces `DummyModuleResolver` at line 139
+  of `executor-core::engine::build_engine`.
+- **Module-shape validation** — `kind = 'module'` source must contain
+  only `fn` declarations, `const` declarations, and `import`
+  statements at top level (no executable expressions). Walks
+  `ast.statements()` via `rhai/internals`. Admin endpoint is the
+  primary gate; the resolver re-runs the check at load time for
+  defense in depth against DB-direct inserts.
+- **Per-module compiled-Module cache** — `LruCache<(AppId, name),
+  (updated_at, Arc<rhai::Module>)>` owned by `Engine`. Invalidated
+  lazily on `updated_at` mismatch. Size via
+  `PICLOUD_MODULE_CACHE_SIZE` (default 512).
+- **Top-level script AST cache** — `LruCache<ScriptId, (updated_at,
+  Arc<rhai::AST>)>` owned by `LocalExecutorClient`. Same staleness
+  semantics. Size via `PICLOUD_SCRIPT_CACHE_SIZE` (default 256).
+- **`ScriptIdentity` + `ExecutorClient::execute_with_identity`** —
+  new method on the trait; default impl forwards to `execute` so
+  `RemoteExecutorClient` (and future transports) keep working.
+  `LocalExecutorClient` overrides it to consult the script cache and
+  pass the resulting `Arc<rhai::AST>` to `Engine::execute_ast`.
+- **`Engine::execute_ast`** — companion to `execute` that takes a
+  pre-compiled AST so callers (the orchestrator) can reuse one
+  compile across many invocations.
+- **Import depth limit** — `Limits::module_import_depth_max`
+  (default 8). Not script-overridable.
+- **Reserved module names** — module-kind scripts cannot be named
+  `log`, `regex`, `random`, `time`, `json`, `base64`, `hex`, `url`,
+  `kv`, `docs`, `dead_letters`, `http`, `files`, `pubsub`, `secrets`,
+  `email`, `users`, `queue`. Defense against author confusion with
+  stdlib namespaces.
+
+### Changed
+
+- **Workspace version**: `1.1.2` → `1.1.3`.
+- **Rhai SDK version**: `1.3` → `1.4` (additive — every v1.3 script
+  still runs unchanged; new surface: `import "<name>" as <alias>;`
+  for endpoint scripts that consume modules in the same app).
+- **Dashboard version**: `0.8.0` → `0.9.0`. Adds kind dropdown on
+  script create + kind badges on the scripts list and detail page.
+- **`Services` bundle** — grows a `modules: Arc<dyn ModuleSource>`
+  field. Constructor signature becomes
+  `Services::new(kv, docs, dead_letters, events, modules)`.
+- **`ScriptValidator` trait** — `validate` now returns
+  `ValidatedScript { imports: Vec<String> }` so the repo can write
+  dep-graph edges in the same transaction as the script row. New
+  `validate_module` method enforces module-shape rules.
+- **Trigger creation tightening** — `POST /api/v1/admin/apps/{id}/triggers/{kv,docs,dead_letter}`
+  now load the target script and reject when (1) it doesn't exist,
+  (2) it belongs to a different app (latent v1.1.1/v1.1.2 gap —
+  closed in v1.1.3), or (3) it is `kind = 'module'`.
+- **Route creation** — `POST /api/v1/admin/scripts/{id}/routes`
+  returns 400 when the target script is `kind = 'module'`.
+
+### Security fix
+
+- **Cross-app trigger target (CVE-class: broken access control).** In
+  v1.1.1 and v1.1.2, `POST /api/v1/admin/apps/{id}/triggers/{kv,docs,dead_letter}`
+  validated only that the caller could manage triggers on `{id}` — it
+  did **not** verify that the target `script_id` belonged to that same
+  app. A member with trigger-management rights on app A could therefore
+  register a trigger in A pointing at a script owned by app B, causing
+  B's script to execute on A's events (a cross-app isolation break).
+  v1.1.3 closes this: every trigger-create handler now loads the target
+  script and rejects it unless `script.app_id == path app_id` (and it is
+  not a module). **Upgrade recommendation:** anyone running a pre-v1.1.3
+  multi-tenant deploy should upgrade and audit existing `triggers` rows
+  for any whose `script_id` resolves to a script in a different `app_id`.
+
+### Migrations
+
+- `0015_scripts_kind.sql` — adds `scripts.kind` with CHECK
+  `IN ('endpoint','module')`, composite index `(app_id, kind)`, and
+  a module-name shape CHECK (`^[a-zA-Z_][a-zA-Z0-9_]{0,63}$`).
+- `0016_script_imports.sql` — adds the dep-graph table with FK
+  CASCADE on both columns, PK `(importer, imported)`, and a
+  reverse-edge index on `imported_script_id`.
+
+### Downgrade caveats
+
+Rolling back v1.1.3 → v1.1.2 with module-kind scripts present
+strands them (no `kind` column means everything looks like an
+endpoint; modules will then succeed as route targets and immediately
+fail to execute meaningfully). Migration `0016_script_imports.sql`
+is safe to drop (the table is auxiliary). `0015_scripts_kind.sql`
+must be reversed by `DROP COLUMN kind` only after manually re-homing
+or deleting module-kind rows.
+
+## v1.1.2 — Documents (unreleased)
+
+`docs::*` SDK — schemaless JSONB document storage with a first-cut
+query DSL — plus `docs:*` triggers as the second concrete kind on the
+v1.1.1 triggers framework. Sets the precedent for the v1.2 query DSL
+expansion and `dead_letters::list`.
+
+### Added
+
+- **Docs store** — `docs` table keyed `(app_id, collection, id)` with
+  JSONB values and a GIN-on-`jsonb_path_ops` index. Rhai SDK exposes
+  the handle pattern:
+  `docs::collection(name).{create,get,find,find_one,update,delete,list}`.
+  Cursor-style pagination on `list`. Cross-app isolation enforced via
+  `cx.app_id` (never script-passed). Document envelope shape returned
+  by reads: `#{ id, data: #{...}, created_at, updated_at }` — explicit
+  metadata + user-data separation (sets precedent for v1.2
+  `dead_letters::list`).
+- **Query DSL (v1.1.2 subset)** — implicit equality at top level
+  (`#{ tier: "gold" }`), operator-object form
+  (`#{ created_at: #{ "$gt": "..." } }`), dotted field paths up to 5
+  levels (`"user.email"`), and operators `$eq`/`$ne`/`$gt`/`$gte`/
+  `$lt`/`$lte`/`$in`. Filter modifiers `$sort` (single field) and
+  `$limit`. Unsupported operators (`$or`, `$regex`, etc.) reject with
+  a clear v1.2-pointer error.
+- **Docs triggers (`docs:*`)** — `docs_trigger_details` table mirrors
+  `kv_trigger_details`. Admin endpoint
+  `POST /api/v1/admin/apps/{id}/triggers/docs` accepts the same DTO
+  shape as the KV endpoint with `ops` of `DocsEventOp` (create /
+  update / delete). Dispatcher routes `OutboxSourceKind::Docs` through
+  the same generic path as KV + dead-letter.
+- **`ctx.event.docs.prev_data`** — change-data-capture surface for
+  docs trigger handlers. `prev_data` carries the document state prior
+  to the mutation (`None` for create), letting handlers see what
+  changed. The repo reads the old row in the same SQL statement as
+  the write so the trigger event has the prior value.
+- **`Capability::AppDocsRead(AppId)`** + `AppDocsWrite(AppId)` —
+  granted to Viewer / Editor respectively in the per-app role table.
+  Same trust shape as KV's `AppKvRead` / `AppKvWrite`.
+
+### Changed
+
+- **Workspace version**: `1.1.1` → `1.1.2`.
+- **Rhai SDK version**: `1.2` → `1.3` (additive — every v1.2 script
+  still runs unchanged; new surfaces: `docs::collection(name).{...}`,
+  `ctx.event.docs` for triggered handlers).
+- **Dashboard version**: `0.7.0` → `0.8.0`. Workspace alignment; no
+  docs-specific UI in v1.1.2 (the dashboard's Rhai-mode hints don't
+  list KV completions either — focused UX pass is a separate task).
+- **`Services` bundle** — grows a `docs: Arc<dyn DocsService>` field.
+  Constructor signature becomes
+  `Services::new(kv, docs, dead_letters, events)`.
+- **Scope mapping**: API keys with `script:read` scope can call
+  `docs::find` / `get` / `list`; `script:write` can call
+  `docs::create` / `update` / `delete`. Same trust shape as KV —
+  honors the seven-scope commitment from v1.1.0.
+
+### Migrations
+
+- `0013_docs.sql` — `docs` table + per-`(app_id, collection)` index +
+  GIN-on-`jsonb_path_ops` index.
+- `0014_docs_triggers.sql` — extends `triggers.kind` and
+  `outbox.source_kind` CHECK constraints to include `'docs'`; adds
+  `docs_trigger_details` table.
+
+### Downgrade caveats
+
+Rolling a deployment back from v1.1.2 → v1.1.1 with `docs`-source
+outbox rows still queued will cause the v1.1.1 dispatcher to fail
+deserialising `TriggerEvent::Docs` (`#[serde(tag = "source")]`
+rejects unknown variants). Drain or delete
+`outbox WHERE source_kind = 'docs'` before downgrading. Trunk-only
+deployments don't hit this.
+
+### Known limitations
+
+- Text-lex comparison for `$gt` / `$gte` / `$lt` / `$lte` is
+  incorrect for unpadded numbers crossing digit-count boundaries
+  (`'10' < '9'` is TRUE under any text collation). Workaround:
+  zero-pad numeric strings. v1.2's advanced query expansion adds
+  numeric-aware operators.
+- Concurrent `update()`s on the same doc may both emit the
+  pre-update `prev_data` (last-writer-wins). Inherited from KV's
+  `set` pattern; documented for forensic-trace use cases.
+- v1.1.2 has no partial-update DSL — scripts that want partial
+  update do `get + modify + update`. Planned for v1.2.
+
 ## v1.1.1 — Storage & Events (unreleased)
 
 The triggers framework — KV store + universal outbox + dispatcher +

CLAUDE.md

@@ -100,7 +100,7 @@ docs/
 
 ## Working Rules
 
-- **Honor the three-service boundary.** Don't reach across `*-core` crates. If `orchestrator-core` needs something from `manager-core`, define a trait in `shared` and inject the impl.
+- **Honor the three-service boundary.** Don't reach across `*-core` crates *for behavior*. If `orchestrator-core` needs to invoke logic from `manager-core`, define a trait in `shared` and inject the impl — keep implementations decoupled. **Transport DTOs are not behavior**: types like `ExecRequest` / `ExecResponse` / `ExecError` represent values produced or consumed across the wire, and depending on the originating crate's type definitions is fine. The bright line is "don't call across crates," not "don't import types." When in doubt: if the imported item is a `struct`/`enum`/`type alias` with no methods (or only data-shape methods), it's a DTO and crossing is fine; if it's a trait, function, or service, define the abstraction in `shared` and inject.
 - **`executor-core` has no Postgres dependency.** Data-plane services (kv, docs, users — v1.1+) come in via injected `ServiceProvider` traits.
 - **Database writes only from `manager-core`.** `orchestrator-core` reads scripts (cached); `executor-core` doesn't touch the DB.
 - **Stateful SDK services use the handle pattern + `SdkCallCx`.** Collection-scoped surfaces look like `kv::collection("x").get(k)`, not `kv::get("x", k)`. Every service trait method takes `&SdkCallCx` and **MUST** derive `app_id` from `cx.app_id` — never trust a script-passed `app_id`. That is the cross-app isolation boundary. See [docs/sdk-shape.md](docs/sdk-shape.md).
@@ -118,6 +118,8 @@ Environment variables consumed by the `picloud` binary:
 | `DATABASE_URL` | — | Required. Postgres connection string. |
 | `PICLOUD_SESSION_TTL_HOURS` | `24` | Sliding-window session lifetime. |
 | `PICLOUD_SANDBOX_MAX_*` | conservative defaults | Per-knob admin ceilings on Rhai sandbox overrides. See `manager-core::sandbox::SandboxCeiling`. |
+| `PICLOUD_FILES_ROOT` | `./data` | Filesystem root for `files::*` blob storage (v1.1.5). Bytes live at `<root>/files/<app_id>/<collection>/<id[0:2]>/<id>`; metadata in Postgres. |
+| `PICLOUD_FILES_MAX_FILE_SIZE_BYTES` | `104857600` (100 MB) | Per-file hard size cap for `files::*` (v1.1.5). Per-app quotas deferred to v1.2. |
 
 ## Out of MVP
 

Cargo.lock

@@ -378,6 +378,28 @@ dependencies = [
  "windows-link",
 ]
 
+[[package]]
+name = "chrono-tz"
+version = "0.9.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "93698b29de5e97ad0ae26447b344c482a7284c737d9ddc5f9e52b74a336671bb"
+dependencies = [
+ "chrono",
+ "chrono-tz-build",
+ "phf",
+]
+
+[[package]]
+name = "chrono-tz-build"
+version = "0.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0c088aee841df9c3041febbb73934cfc39708749bf96dc827e3359cd39ef11b1"
+dependencies = [
+ "parse-zoneinfo",
+ "phf",
+ "phf_codegen",
+]
+
 [[package]]
 name = "clap"
 version = "4.6.1"
@@ -499,6 +521,17 @@ version = "2.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "217698eaf96b4a3f0bc4f3662aaa55bdf913cd54d7204591faa790070c6d0853"
 
+[[package]]
+name = "cron"
+version = "0.12.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6f8c3e73077b4b4a6ab1ea5047c37c57aee77657bc8ecd6f29b0af082d0b0c07"
+dependencies = [
+ "chrono",
+ "nom",
+ "once_cell",
+]
+
 [[package]]
 name = "crossbeam-queue"
 version = "0.3.12"
@@ -1274,6 +1307,15 @@ version = "0.4.29"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "5e5032e24019045c762d3c0f28f5b6b8bbf38563a65908389bf7978758920897"
 
+[[package]]
+name = "lru"
+version = "0.12.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "234cf4f4a04dc1f57e24b96cc0cd600cf2af460d4161ac5ecdd0af8e1f3b2a38"
+dependencies = [
+ "hashbrown 0.15.5",
+]
+
 [[package]]
 name = "lru-slab"
 version = "0.1.2"
@@ -1317,6 +1359,12 @@ version = "0.3.17"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6877bb514081ee2a7ff5ef9de3281f14a4dd4bceac4c09388074a6b5df8a139a"
 
+[[package]]
+name = "minimal-lexical"
+version = "0.2.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
+
 [[package]]
 name = "mio"
 version = "1.2.0"
@@ -1337,6 +1385,16 @@ dependencies = [
  "spin 0.5.2",
 ]
 
+[[package]]
+name = "nom"
+version = "7.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a"
+dependencies = [
+ "memchr",
+ "minimal-lexical",
+]
+
 [[package]]
 name = "normalize-line-endings"
 version = "0.3.0"
@@ -1454,6 +1512,15 @@ dependencies = [
  "windows-link",
 ]
 
+[[package]]
+name = "parse-zoneinfo"
+version = "0.3.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1f2a05b18d44e2957b88f96ba460715e295bc1d7510468a2f3d3b44535d26c24"
+dependencies = [
+ "regex",
+]
+
 [[package]]
 name = "password-hash"
 version = "0.5.0"
@@ -1503,9 +1570,47 @@ version = "2.3.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9b4f627cb1b25917193a259e49bdad08f671f8d9708acfd5fe0a8c1455d87220"
 
+[[package]]
+name = "phf"
+version = "0.11.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1fd6780a80ae0c52cc120a26a1a42c1ae51b247a253e4e06113d23d2c2edd078"
+dependencies = [
+ "phf_shared",
+]
+
+[[package]]
+name = "phf_codegen"
+version = "0.11.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "aef8048c789fa5e851558d709946d6d79a8ff88c0440c587967f8e94bfb1216a"
+dependencies = [
+ "phf_generator",
+ "phf_shared",
+]
+
+[[package]]
+name = "phf_generator"
+version = "0.11.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3c80231409c20246a13fddb31776fb942c38553c51e871f8cbd687a4cfb5843d"
+dependencies = [
+ "phf_shared",
+ "rand 0.8.6",
+]
+
+[[package]]
+name = "phf_shared"
+version = "0.11.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "67eabc2ef2a60eb7faa00097bd1ffdb5bd28e62bf39990626a582201b7a754e5"
+dependencies = [
+ "siphasher",
+]
+
 [[package]]
 name = "picloud"
-version = "1.1.1"
+version = "1.1.5"
 dependencies = [
  "anyhow",
  "async-trait",
@@ -1531,7 +1636,7 @@ dependencies = [
 
 [[package]]
 name = "picloud-cli"
-version = "1.1.1"
+version = "1.1.5"
 dependencies = [
  "anyhow",
  "assert_cmd",
@@ -1552,7 +1657,7 @@ dependencies = [
 
 [[package]]
 name = "picloud-executor"
-version = "1.1.1"
+version = "1.1.5"
 dependencies = [
  "anyhow",
  "picloud-executor-core",
@@ -1564,12 +1669,13 @@ dependencies = [
 
 [[package]]
 name = "picloud-executor-core"
-version = "1.1.1"
+version = "1.1.5"
 dependencies = [
  "async-trait",
  "base64",
  "chrono",
  "hex",
+ "lru",
  "percent-encoding",
  "picloud-shared",
  "rand 0.8.6",
@@ -1580,12 +1686,14 @@ dependencies = [
  "thiserror 1.0.69",
  "tokio",
  "tracing",
+ "tracing-subscriber",
+ "url",
  "uuid",
 ]
 
 [[package]]
 name = "picloud-manager"
-version = "1.1.1"
+version = "1.1.5"
 dependencies = [
  "anyhow",
  "picloud-manager-core",
@@ -1597,18 +1705,21 @@ dependencies = [
 
 [[package]]
 name = "picloud-manager-core"
-version = "1.1.1"
+version = "1.1.5"
 dependencies = [
  "argon2",
  "async-trait",
  "axum",
  "base64",
  "chrono",
+ "chrono-tz",
+ "cron",
  "data-encoding",
  "picloud-executor-core",
  "picloud-orchestrator-core",
  "picloud-shared",
  "rand 0.8.6",
+ "reqwest",
  "serde",
  "serde_json",
  "sha2",
@@ -1622,7 +1733,7 @@ dependencies = [
 
 [[package]]
 name = "picloud-orchestrator"
-version = "1.1.1"
+version = "1.1.5"
 dependencies = [
  "anyhow",
  "picloud-orchestrator-core",
@@ -1634,14 +1745,16 @@ dependencies = [
 
 [[package]]
 name = "picloud-orchestrator-core"
-version = "1.1.1"
+version = "1.1.5"
 dependencies = [
  "async-trait",
  "axum",
  "chrono",
+ "lru",
  "picloud-executor-core",
  "picloud-shared",
  "reqwest",
+ "rhai",
  "serde",
  "serde_json",
  "thiserror 1.0.69",
@@ -1653,7 +1766,7 @@ dependencies = [
 
 [[package]]
 name = "picloud-shared"
-version = "1.1.1"
+version = "1.1.5"
 dependencies = [
  "async-trait",
  "chrono",
@@ -2356,6 +2469,12 @@ dependencies = [
  "rand_core 0.6.4",
 ]
 
+[[package]]
+name = "siphasher"
+version = "1.0.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8ee5873ec9cce0195efcb7a4e9507a04cd49aec9c83d0389df45b1ef7ba2e649"
+
 [[package]]
 name = "slab"
 version = "0.4.12"

Cargo.toml

@@ -13,7 +13,7 @@ members = [
 ]
 
 [workspace.package]
-version = "1.1.1"
+version = "1.1.5"
 edition = "2021"
 rust-version = "1.92"
 license = "MIT OR Apache-2.0"
@@ -47,12 +47,16 @@ tracing-subscriber = { version = "0.3", features = ["env-filter", "json"] }
 # IDs + time
 uuid = { version = "1", features = ["v4", "serde"] }
 chrono = { version = "0.4", features = ["serde"] }
+# Cron schedule parsing (v1.1.4 cron triggers) + IANA timezone resolution.
+chrono-tz = "0.9"
+cron = "0.12"
 
 # Async traits
 async-trait = "0.1"
 
-# Rhai scripting
-rhai = { version = "1.19", features = ["sync", "serde"] }
+# Rhai scripting. Pinned exactly (`=1.24`) because the `internals`
+# feature surface is not semver-stable — future bumps must be deliberate.
+rhai = { version = "=1.24", features = ["sync", "serde"] }
 
 # Postgres (manager-core only — others stay DB-free)
 sqlx = { version = "0.8", features = ["runtime-tokio-rustls", "postgres", "uuid", "chrono", "json", "macros", "migrate"] }
@@ -80,6 +84,10 @@ regex = "1"
 hex = "0.4"
 percent-encoding = "2"
 
+# LRU caches (v1.1.3 — top-level script AST cache in orchestrator-core +
+# per-module compiled-module cache in executor-core).
+lru = "0.12"
+
 [workspace.lints.rust]
 unsafe_code = "forbid"
 

HANDBACK.md

@@ -1,340 +1,127 @@
-# v1.1.1 Implementation HANDBACK
+# HANDBACK — v1.1.5 Files & Pub/Sub
 
-## 1. Branch + commit count
+## §1 Branch + commits
 
-- Branch: `feat/v1.1.1-storage-and-events`
-- Base: `main`
-- 11 commits ahead of `main`. Branch is **not pushed**, **not merged**.
+- **Branch:** `feat/v1.1.5-files-pubsub` (off `main`). Not pushed, not merged, no PR.
+- **Commits:** the two-feature split decided in planning + a finalize commit; HANDBACK is the 4th (docs):
+  1. `6e132b6 feat(v1.1.5): files SDK + files:* triggers`
+  2. `834c787 feat(v1.1.5): pubsub::publish_durable SDK + pubsub:* triggers`
+  3. `4595db7 chore(v1.1.5): version bumps, CI workflow, schema-snapshot un-ignore`
+  4. `docs(v1.1.5): handback report` (this file)
 
-```
-66b661f chore(release): bump workspace to v1.1.1 + CHANGELOG
-6b7ff78 feat(v1.1.1-gc): dead-letter + abandoned-executions retention sweepers
-1795dfc feat(v1.1.1-dead-letters): dashboard badge + list view
-20f1b5e feat(v1.1.1-dead-letters): service + Rhai SDK + admin endpoints
-77b2cb5 feat(v1.1.1-routes): outbox-routed sync HTTP + dispatch_mode=async
-6a2971a feat(v1.1.1-dispatcher): dispatcher loop + retry + depth limit + outbox emitter
-2e92691 feat(v1.1.1-triggers): trigger CRUD admin endpoints
-545d863 feat(v1.1.1-triggers): triggers + outbox schema + repos
-6b99f74 feat(v1.1.1-kv): Rhai kv:: SDK module + ctx.event wiring
-434fb63 feat(v1.1.1-kv): migrations + KvService trait + Postgres impl
-1efb350 docs(v1.1.x): resolve in-flight decisions as Decided 2026-06-01
-```
+Each of commits 1–3 is independently green (fmt + clippy + `cargo test --workspace`). Shared files (Cargo deps, `Services` bundle, `version.rs`, dispatcher arm, authz enum, CHANGELOG) are touched in both feature commits as planned — additive only, so commit 1 compiles green with the `AppPubsubPublish` capability and the dashboard `'pubsub'` type union present-but-unused until commit 2.
 
-The first commit (`1efb350`) absorbed working-tree edits to
-`docs/v1.1.x-design-notes.md` that turned the "in-flight" 20 open
-calls into "Decided 2026-06-01" entries. Those were on the working
-tree at branch creation; folding them into the v1.1.1 branch keeps
-the design rationale colocated with the implementation.
+## §2 Scope coverage
 
-## 2. Scope coverage (Done / Partial / Skipped)
-
-| Scope item | Status | Notes |
+| Brief item | Status | Notes |
 |---|---|---|
-| **1. KV store** | Done | Migration 0007, `KvService` trait in shared, `KvServiceImpl` + `PostgresKvRepo` in manager-core, Rhai `kv::collection(name).{get,set,has,delete,list}` bridge, cursor pagination, empty-collection rejection, script-as-gate authz. |
-| **2. Triggers framework — Layout E** | Done | Migrations 0008 (`triggers` + `kv_trigger_details` + `dead_letter_trigger_details`), `TriggerRepo` + `PostgresTriggerRepo`, CRUD admin endpoints. `registered_by_principal` column captured + threaded into the dispatcher. Depth-limit enforced in the dispatcher (default 8). |
-| **3. Universal outbox + dispatcher** | Done | Migration 0009 (`outbox`), `OutboxRepo` + `PostgresOutboxRepo`, `Dispatcher` tokio task. Polls every 100ms, claims 8 rows/tick via `FOR UPDATE SKIP LOCKED`, gate-bounds dispatch, retries with backoff+jitter, dead-letters on exhaustion, late-completion → `abandoned_executions`. |
-| **4. NATS-style sync HTTP** | Done | `InboxRegistry` in orchestrator-core (in-process `Mutex<HashMap<Uuid, oneshot::Sender>>`), `InboxResolver` trait in shared. Orchestrator sync-route path registers receiver, writes outbox row with `reply_to`, awaits with timeout = script.timeout + 2s buffer. Status mapping per design notes §3 (422/502/503/504/507/500). |
-| **5. `dispatch_mode: async` HTTP routes** | Done | Migration 0012 adds the column (default `sync`). `DispatchMode` enum in shared. Route admin payload + RouteRepository serialize it. Compiled routes carry it; the matcher returns it in `Matched`. Orchestrator branches: async → outbox + 202; sync → outbox + inbox. |
-| **6. Dead letters** | Done | Migration 0010 (`dead_letters`), `DeadLetterRepo` + `DeadLetterService` + `PostgresDeadLetterService`. Rhai `dead_letters::{replay,resolve}` + admin endpoints (`GET /count`, `GET /`, `GET /{id}`, `POST /{id}/replay`, `POST /{id}/resolve`). `Capability::AppDeadLetterManage(AppId)` enforced. List intentionally NOT shipped (deferred to v1.2). Recursion-stop rule (handler-failure annotates original DL as `handler_failed`) implemented in the dispatcher. |
-| **7. Abandoned executions** | Done | Migration 0011, `AbandonedRepo` + `PostgresAbandonedRepo`, dispatcher writes a row on dropped-receiver inbox delivery. Metric path reserved (`TODO(metrics)` markers in dispatcher.rs). |
-| **8. Retry policy defaults** | Done | `TriggerConfig::from_env` (new module). Env vars: `PICLOUD_MAX_TRIGGER_DEPTH`, `PICLOUD_TRIGGER_RETRY_{MAX_ATTEMPTS,BACKOFF,BASE_MS,JITTER_PCT}`, `PICLOUD_DEAD_LETTER_RETENTION_DAYS`, `PICLOUD_ABANDONED_EXECUTIONS_RETENTION_DAYS`. Per-trigger overrides applied at trigger-creation time. |
-| **9. `ctx.event` for triggered scripts** | Done | `TriggerEvent` enum in shared (KV / DeadLetter variants), `SdkCallCx.event: Option<TriggerEvent>` + `is_dead_letter_handler: bool`. `engine.rs::build_ctx_map` flattens the event into `ctx.event` for triggered handlers; direct invocations leave the key absent. Shape matches design notes §4 (KV with op + collection + key + value; dead_letter with original + attempts + last_error + ids + timestamps). |
-| **10. Dashboard surface** | Done | Per-app red badge with unresolved count on apps list + per-app detail page. New `apps/[slug]/dead-letters/+page.svelte` list view with all design-notes-mandated columns + Replay + Mark resolved actions + expandable row detail. svelte-check passes (369 files, 0 errors, 0 warnings). |
-| **11. Workspace version bump** | Done | Workspace `1.1.0` → `1.1.1`, SDK `1.1` → `1.2`, dashboard `0.6.0` → `0.7.0`. CHANGELOG.md created at repo root. |
+| §1 `files::*` SDK | ✅ | `create/head/get/update/delete/list`, blob in/out, metadata maps, throw-vs-`()` convention. |
+| §1 migration 0018_files.sql | ✅ | metadata table + `idx_files_app_collection`. Bytes on disk, never in PG. |
+| §1 atomic writes/deletes, checksum, size+name+type caps, authz, events | ✅ | See §3. |
+| §2 `files:*` trigger (Layout-E, 0019) | ✅ | widen 2 CHECKs + `files_trigger_details`; `TriggerEvent::Files` (metadata only); admin `POST /triggers/files`; `emit_files`; dispatcher arm. |
+| §3 `pubsub::publish_durable` SDK | ✅ | publish-time transactional fan-out; topic matching in Rust; succeed-silently on no match. |
+| §4 `pubsub:*` trigger (Layout-E, 0020) | ✅ | widen 2 CHECKs + `pubsub_trigger_details` + partial index; `TriggerEvent::Pubsub`; admin `POST /triggers/pubsub`; dispatcher arm. |
+| §5 dashboard Files view | ✅ | `apps/[slug]/files/+page.svelte` (list per collection, per-row delete w/ confirm). Backed by a new admin files API (§7.2). |
+| §5 dashboard Pub/Sub trigger form | ✅ | added to the Triggers tab beside Cron; trigger-list renders files + pubsub. `npm run check` clean. |
+| §6 schema_snapshot CI follow-up | ✅ | §6b skip-when-absent + un-ignore; §6a new `.github/workflows/ci.yml`. See §5. |
+| §7 version bumps | ✅ | workspace 1.1.4→1.1.5, SDK 1.5→1.6, dashboard 0.10.0→0.11.0, CHANGELOG, CLAUDE.md env table. |
+| §8 tests | ⚠️ | 63 new tests (target 70–90). Every *named* critical test covered; gap is the dispatcher end-to-end DB test (see §9.2). |
 
-## 3. Key implementation decisions / deviations
+## §3 Files implementation notes
 
-### Outbox column set (deferred to implementation per design notes §2)
-Chose:
-- `script_id` denormalized — dispatcher resolves the target without
-  re-joining for the common path.
-- `trigger_id` polymorphic (no DB FK) — references `triggers.id` for
-  `source_kind IN {kv, dead_letter}`, `routes.id` for
-  `source_kind = 'http'`. Discrimination in Rust at dispatch time.
-- `claimed_by TEXT` — pid-based for MVP; cluster mode can use any
-  identifier without schema change.
-- `trigger_depth` + `root_execution_id` denormalized so the
-  dispatcher rebuilds `ExecRequest` without joining back to the
-  originating execution log.
-- No explicit `is_dead_letter_handler` column — dispatcher infers
-  from the trigger's `kind` field at dispatch time.
+**Service layering** (`FilesServiceImpl`, manager-core): validate collection (empty + traversal) → script-as-gate authz (`AppFilesRead`/`AppFilesWrite`, skipped when `cx.principal` is `None`) → field/size-cap validation → repo call keyed by `cx.app_id` → best-effort `ServiceEvent` emit. `executor-core` has **no** Postgres or filesystem dependency — both traits live in `picloud-shared`, the impl in manager-core.
 
-### KV pagination
-- **Cursor-style**, opaque base64-encoded last-key.
-- Page-size cap of 1000 with default 100 (enforced in repo).
-- Documented in `crates/shared/src/kv.rs` and the SDK function
-  comment.
+**Atomic-write protocol** (`write_atomic_at`, a free fn so it's unit-testable without a pool):
+1. Validate collection path-safety (defensive — already enforced at the SDK boundary).
+2. `create_dir_all` the shard dir `<root>/files/<app_id>/<collection>/<id[0:2]>/<id>` with `0o700` (Unix `DirBuilderExt::mode`).
+3. SHA-256 the in-memory bytes (single pass — never re-reads the file) while writing to `<final>.tmp.<pid>-<atomic-counter>`.
+4. `sync_all()` the temp file.
+5. `rename(tmp, final)` — atomic on POSIX.
+6. `sync_all()` the parent dir (rename durability).
+7. INSERT/UPDATE the DB row.
 
-### KV TTL
-- Blueprint §8.1 reserved an `expires_at` column. v1.1.1 design notes
-  don't surface TTL through the SDK (`set(k,v)` has no TTL argument)
-  so the column is **omitted from migration 0007**. Adding it later
-  is a non-breaking forward migration. Recorded in CHANGELOG as a
-  deferred item.
+Rollback per step: crash in 1–5 → orphan `*.tmp.*` (never read; the pid+counter suffix avoids collisions); crash in 5–7 → bytes with no row, **never reachable via the SDK** because every read starts from the row. `update` reads the prior row first (existence + CDC `prev`), writes new bytes, then UPDATEs.
 
-### Authz scope mapping (seven-scope commitment)
-The four new capabilities map onto existing scopes — **no new scope
-variants** to honour the `Scope` enum's "exactly seven values"
-contract (`crates/shared/src/auth.rs:103`):
+**Atomic-delete protocol** (`FsFilesRepo::delete`): `SELECT … FOR UPDATE` + `DELETE` in one transaction → commit → `unlink` outside the tx. Unlink failure leaves an orphan (logged at warn); failure before commit changes nothing. Returns the deleted metadata so the service can emit.
 
-| Capability | Scope |
-|---|---|
-| `AppKvRead` | `script:read` |
-| `AppKvWrite` | `script:write` |
-| `AppManageTriggers` | `app:admin` |
-| `AppDeadLetterManage` | `app:admin` |
+**Path-traversal validation:** `picloud_shared::validate_files_collection` rejects empty / `/` / `\` / `..` / NUL at the SDK boundary; `FsFilesRepo::guard_collection` repeats it before any fs op. UUID ids can't produce traversal (verified defensively).
 
-`role_satisfies` grants `AppKvRead` at the Viewer role, `AppKvWrite`
-at Editor, and both trigger / DL caps at AppAdmin.
+**Per-call SHA-256:** computed once over the in-memory `Vec<u8>` during the write (`sha2::Sha256`), hex-lowercased, stored on the row. The file is never re-read to hash. Known-vector tests pin `SHA-256("abc")` and `SHA-256("")`.
 
-### Script-as-gate authz for SDK calls
-- `KvServiceImpl` runs `authz::require` only when
-  `cx.principal.is_some()`. Anonymous public-HTTP scripts (the
-  common case for public routes) bypass the cap check.
-- Cross-app isolation is **independent** of this — enforced by
-  `cx.app_id` being the only source of `app_id` on every query.
-- `PostgresDeadLetterService::{replay,resolve}` keeps a hard
-  `require` (no `if let Some`) — managing dead letters is an admin
-  act per design notes §4. Public scripts with `principal: None`
-  fail the check, which is correct.
+**Checksum-on-get:** `get` reads the file, re-hashes, compares to the stored checksum. Mismatch (or missing bytes while the row persists) → `FilesError::Corrupted`, logged at error level with the path, **no auto-delete**. To scripts this surfaces as a thrown Rhai error `"files: file content corrupted (checksum mismatch)"`.
 
-### Trait split: `OutboxRepo` vs `OutboxWriter`
-orchestrator-core can't depend on manager-core (would invert the
-dependency arrow). Defined a small `OutboxWriter` trait in
-`picloud-shared` with a single `enqueue_http` method.
-`PostgresOutboxRepo` implements both `OutboxRepo` (dispatcher
-surface) and `OutboxWriter` (orchestrator surface); the picloud
-binary clones one concrete Arc into both trait views — mirrors the
-existing `members_concrete` / `AuthzRepo` pattern.
+## §4 Pub/Sub implementation notes
 
-### `InboxResolver` lives in shared, `InboxRegistry` in orchestrator-core
-Same split rationale — the dispatcher (manager-core) only depends on
-the trait, while the in-process impl lives next to its consumer.
-Cluster mode (v1.3+) swaps the impl for `LISTEN/NOTIFY` behind the
-unchanged trait.
+**Fan-out-at-publish-time, transactional** (`PostgresPubsubRepo::fan_out_publish`): one transaction — `SELECT` all enabled pubsub triggers for the app (joined to `pubsub_trigger_details`), filter by `topic_matches` in Rust, INSERT one `outbox` row (`source_kind='pubsub'`) per survivor, commit once. A mid-fan-out failure rolls back every row (no half-fan-out). Each delivery row then retries/dead-letters independently through the unchanged dispatcher (its trigger arm just gained `| OutboxSourceKind::Pubsub`).
 
-### manager-core now depends on executor-core
-Previously manager-core only depended on orchestrator-core. The
-dispatcher needs `ExecRequest`/`ExecResponse`/`ExecError`/
-`InvocationType` from `executor-core` to build invocation
-descriptors. This is the transport DTO interpretation of the
-working-rules "don't reach across `*-core` crates" — DTOs are fine,
-behaviour is the bright line.
+**Topic pattern matching** runs in Rust (`picloud_shared::topic_matches`), not SQL: `"*"` → all; `"<prefix>.*"` → `starts_with("<prefix>.")`; otherwise exact. `validate_topic_pattern` (used at trigger creation in the admin endpoint and defensively in the repo) accepts only `*` / `<prefix>.*` / no-star-exact, rejecting `*.created`, `**`, `a.*.b`, `user.*x`, etc. with `"unsupported pubsub topic pattern: …"`.
 
-### Sync HTTP via outbox is the default for the user-routes path
-The orchestrator's user-route handler is fully on the NATS-style
-path now — every sync HTTP request writes to the outbox and awaits
-inbox delivery. Adds ~2-5ms per request per design notes §3 latency
-budget. `/api/v1/execute/{id}` (the admin/dev bypass) still calls
-the executor directly since it doesn't need the unified
-observability — kept for simplicity and admin tooling speed.
+**No matching trigger → the publish succeeds, zero outbox rows** (the design-notes-preferred succeed-silently). `published_at` is stamped manager-side (`Utc::now()`) so every delivery agrees on one instant. `ctx.event.pubsub = #{ topic, message, published_at }`, `ctx.event.op = "publish"`.
 
-### Trigger-depth check is on the outbox row, not in the executor
-Dispatcher rejects depth-exceeded rows **before** trying to
-execute. The `cx.trigger_depth` field is informational on the
-executor side. Rejection writes a log + (reserved) metric and
-deletes the row — no DL, per design notes §4.
+There is **no `list_matching_pubsub` on `TriggerRepo`** — pubsub publishes directly (it's not a `ServiceEvent`), so the fan-out SELECT lives in `pubsub_repo`, not the `OutboxEventEmitter`. This is the one structural asymmetry vs files/kv/docs, intentional per the publish-time-fan-out decision.
 
-## 4. Tests added
+## §5 CI follow-up (§6) status
 
-### Unit tests (no DB required)
-- `manager-core::kv_service::tests` (10 tests) — round-trip, missing
-  key returns None, `has` predicate, `delete` was-present,
-  empty-collection rejection, **cross-app isolation**, anonymous-cx
-  skips authz, authed-cx-with-no-role is Forbidden, owner-can-write,
-  cursor pagination via in-memory KvRepo + denying authz repo.
-- `manager-core::trigger_config::tests` (2 tests) — conservative
-  defaults, backoff round-trips.
-- `manager-core::trigger_repo::tests` (1 test) — `collection_matches`
-  glob behaviour (`*`, `prefix:*`, exact).
-- `manager-core::dispatcher::tests` (5 tests) — exponential / linear /
-  constant backoff math, jitter within bounds, ExecError →
-  InboxFailureKind classification, failure-kind → status-code mapping.
-- `manager-core::abandoned_repo::tests` (2 tests) — truncate
-  char-boundary safety.
-- `manager-core::triggers_api::tests` (5 tests) — unknown-app 404,
-  member-without-role 403, default fallback for retry settings,
-  empty-glob rejection, cross-app delete is treated as not-found.
-- `orchestrator-core::inbox::tests` (4 tests) — register/deliver
-  round-trip, unknown-id is Abandoned, dropped receiver is
-  Abandoned, explicit cancel removes sender.
-- `executor-core::engine::tests` (3 new) — `ctx.event` absent for
-  direct invocations, KV insert shape matches design notes,
-  KV delete has unit value.
-- `executor-core::sdk_kv` integration suite (7 tests) — runs a real
-  Rhai engine under `spawn_blocking` against an in-memory
-  `KvService` impl. Covers handle pattern, round-trip, unit-on-
-  missing, has predicate, delete-was-present, empty-collection
-  throws, cursor pagination, **cross-app isolation through the
-  bridge**.
+- **Pre-existing CI:** none (no `.github/`, no `.gitlab-ci.yml`).
+- **§6a (added):** `.github/workflows/ci.yml` — a `rust` job with a `postgres:15` service (`DATABASE_URL=postgres://picloud:picloud@localhost:5432/picloud`) running `cargo fmt --all -- --check`, `cargo clippy --all-targets --all-features -- -D warnings`, `cargo test --workspace`; a separate `dashboard` job running `npm ci` + `npm run check`.
+- **§6b (done):** `schema_snapshot.rs` is no longer `#[ignore]`'d. Reworked from `#[sqlx::test]` to `#[tokio::test]` that **skips cleanly when `DATABASE_URL` is unset** (chosen over fail-loud so `cargo test --workspace` stays green locally) and otherwise connects, runs `sqlx::migrate!`, and dumps. Golden `expected_schema.txt` re-blessed (now contains `files`, `files_trigger_details`, `pubsub_trigger_details`, both widened CHECKs, `idx_files_app_collection`, `idx_triggers_app_pubsub_enabled`, and migrations 0018–0020).
+  - **Tradeoff (documented):** the non-`sqlx::test` path applies migrations against the `DATABASE_URL` database directly rather than an isolated throwaway DB. Migrations are forward-only/idempotent and CI's Postgres is fresh, so the structural dump is identical; locally it will also apply 0018–0020 to whatever DB you point at.
 
-**Total: 47 new tests across the workspace.** Workspace test counts
-after v1.1.1: 63 manager-core / 56 orchestrator-core / 17
-executor-core engine / 7 sdk_kv / 30 sdk_contract / 43 stdlib /
-21 picloud / 6 shared.
+## §6 Schema decisions beyond the brief
 
-### Intentionally untested
-- DB-backed integration tests for the full dispatcher loop, KV→
-  trigger→DL retry chain, sync HTTP via outbox round-trip,
-  recursion-stop end-to-end. These need a real Postgres harness;
-  the reviewer runs them via the manual smoke flow below.
-- Postgres-specific repo behaviour (sqlx query correctness). The
-  repos compile and run against the schema, but no integration
-  test crate spins up a DB in this branch — same pattern as v1.1.0
-  (see existing `ignored, needs DATABASE_URL` test markers).
+- `files` table is verbatim from the brief. `files_trigger_details` / `pubsub_trigger_details` mirror `kv_trigger_details` / `cron_trigger_details`.
+- `pubsub_trigger_details` has no `ops` column (a publish has a single implicit op) — only `topic_pattern`.
+- `idx_triggers_app_pubsub_enabled` is the third partial index of its kind (per the brief's note); deliberate duplication.
 
-## 5. Open questions for the reviewer
+## §7 Decisions beyond the brief (every prompt-default deviation)
 
-1. **Outbox `claimed_at` clearing on success.** The dispatcher
-   `delete`s the outbox row after success / DL. For failures it
-   reschedules (which sets `claimed_at = NULL`). Both flows are
-   correct, but if you imagine a crash between the executor return
-   and the row update, the row stays claimed forever. Cluster mode
-   should add a periodic "unstick stale claims" sweep. Not in
-   v1.1.1 scope but worth surfacing.
-2. **Sync HTTP overhead.** Every sync HTTP request now goes through
-   the outbox (write + dispatcher pickup + inbox delivery).
-   Measured overhead expected ~2-5ms per design notes §3. No
-   benchmarking yet — recommend the reviewer pick a representative
-   script and compare 95p latency vs v1.1.0 if performance matters.
-3. **HTTP outbox rows don't run as a principal.** The orchestrator's
-   public HTTP path has no authenticated user; the
-   `origin_principal` field on the outbox row is forensic. The
-   resulting `ExecRequest.principal = None`, so the script runs
-   anonymously — matches direct execution. If you'd prefer
-   triggered-from-HTTP scripts to inherit a derived principal
-   (e.g. the route's app's owner), that's an additive change.
-4. **Dispatcher uses `ASYNC_EXEC_TIMEOUT = 300s` for async rows.**
-   Async dispatches don't have a script-level timeout (no
-   originating HTTP request to bound). Picked the same platform
-   cap as `LocalExecutorClient`. If async needs a different cap,
-   easy to thread through `TriggerConfig`.
-5. **Dispatcher tick cadence is 100ms.** Bounded enough that
-   fan-out feels instant; loose enough that an idle process
-   doesn't burn cycles. If the reviewer wants tighter latency,
-   bump to 50ms or use `LISTEN/NOTIFY` for wake-up (v1.3+ work).
-6. **CHANGELOG.md is new.** Followed the rest of the repo's
-   convention from git log (release commits + design-notes
-   references). If a different format is preferred, easy to swap.
+1. **Empty blob treated as a missing `data` field.** `NewFile::validate` / `FileUpdate::validate` reject 0-byte `data` with `FilesError::MissingField("data")`. The brief lists `data` as required and tests "missing … data"; the cleanest testable interpretation at the service layer is "empty == missing". Consequence: v1.1.5 cannot store an intentionally-empty file. Easy to relax later.
+2. **Admin files REST API added** (`files_api.rs`: `GET /apps/{id}/files?collection=…`, `DELETE /apps/{id}/files/{collection}/{file_id}`). The brief's §5 dashboard needs a backend but didn't spell out admin endpoints; I added a minimal one mirroring `triggers_api`'s direct-repo + capability pattern (`AppFilesRead` for list, `AppFilesWrite` for delete).
+3. **Admin file delete does NOT emit a `files:delete` trigger event.** It's an operator cleanup action, not a script mutation, so it goes straight to the repo. SDK deletes still emit. Flagging because "every successful mutation emits" could be read to include admin deletes.
+4. **Files `list` bridge accepts both positional and map forms** — `list()`, `list(cursor)`, `list(cursor, limit)`, and `list(#{ cursor, limit })` (the map form the brief's example used). Additive convenience.
+5. **Files collection-glob semantics reuse the existing `collection_matches`** (`*` / `foo*` prefix / exact), identical to kv/docs. The brief mentioned a `"prefix:*"` form in one spot; I kept parity with the established kv/docs matcher rather than introduce a new glob dialect.
+6. **schema_snapshot runs against the live `DATABASE_URL` DB** rather than an isolated temp DB (see §5).
+7. **Orphan sweep deferred to v1.1.6+** — confirmed with the user during planning (the brief's recommended default). No `*.tmp.*` sweeper daemon shipped.
 
-## 6. Deferred to later releases
+## §8 How to verify locally — attestation (fresh run on HEAD `4595db7`)
 
-- `dead_letters::list(filter)` Rhai SDK — design notes §4 defers
-  to v1.2 to align with `docs::find()` query DSL.
-- KV TTL (`set(k, v, ttl_secs)`) — blueprint reserved it; v1.1.1
-  SDK doesn't surface it. Forward-compat (no schema cost).
-- Auto-disable of triggers whose script was deleted — design notes
-  §4 says current handling is metric+log; auto-disable is v1.2.
-- Per-app dead-letter retention — design notes §4 says env-only in
-  v1.1.1.
-- Metrics counter emit for `picloud_trigger_depth_exceeded`,
-  `picloud_dead_letter_handler_failures`,
-  `picloud_abandoned_executions_total`. Code paths log the
-  occurrences with `tracing::warn`/`error`; the actual
-  counter-emit code is a `TODO(metrics)` comment in the
-  dispatcher. Metrics surface is v1.1.7+ per the roadmap.
-- DB-backed integration tests for the dispatcher loop (see §4
-  intentionally-untested).
-- Sync HTTP performance benchmarks comparing v1.1.0 direct path vs
-  v1.1.1 outbox path.
-
-## 7. How to verify locally
-
-### Static checks (all green on this branch)
-```sh
-cargo fmt --all -- --check
-cargo clippy --all-targets --all-features -- -D warnings
-cargo test --workspace
-cd dashboard && npm run check && npm run build
+```
+cargo fmt --all -- --check                                   → exit 0
+cargo clippy --all-targets --all-features -- -D warnings     → exit 0
+cargo test --workspace                                       → 491 passed, 0 failed (exit 0)
+                                                                (schema_snapshot skips cleanly with no DATABASE_URL)
+cd dashboard && npm run check                                → 0 errors, 0 warnings (exit 0)
 ```
 
-### Migration integrity
-```sh
-docker compose down -v && docker compose up -d postgres
-cargo run -p picloud      # applies 0001..0012 from empty
+With a live Postgres (the schema guardrail actually verifies the schema):
 ```
-Then start from `main` (v1.1.0 schema state) and switch to this
-branch; restart `picloud` to apply 0007..0012 on top.
-
-### Manual end-to-end smoke (reviewer should run)
-```sh
-docker compose up -d
-# 1. Bootstrap an owner user via the existing flow + create app A.
-# 2. Create a script in A whose body is:  throw "boom"
-# 3. POST /api/v1/admin/apps/{A}/triggers/kv with
-#    {"script_id": "<broken>", "collection_glob": "*", "ops": ["insert"]}
-# 4. From another script (or a public HTTP route):
-#    kv::collection("widgets").set("k1", #{n:1})
-# 5. Wait ~7 seconds (3 attempts × ~1/2/4s backoff with ±20% jitter).
-# 6. Open the dashboard at /admin.
-# 7. Apps list shows a red "1" badge next to app A.
-# 8. Click into app A → "Dead letters" tab link → row visible.
-# 9. Click row → full payload + error history.
-# 10. Click "Replay" → row marks resolution='replayed', new outbox
-#     row written, dispatcher re-runs the handler (fails again,
-#     produces a NEW DL row).
-# 11. Click "Mark resolved" on the original DL → resolution='ignored'.
+DATABASE_URL=postgres://picloud:picloud@127.0.0.1:15432/picloud \
+  cargo test -p picloud-manager-core --test schema_snapshot   → test result: ok. 1 passed
 ```
+Migrations 0018–0020 applied cleanly on top of the existing v1.1.4 dev DB during the re-bless — the same `sqlx::migrate!` replay CI runs on a fresh Postgres.
 
-### Async route smoke
-```sh
-# Create a route via POST /api/v1/admin/scripts/{id}/routes with
-# {"host_kind":"any","path_kind":"exact","path":"/work","dispatch_mode":"async"}
-curl -X POST -d '{"work":"thing"}' http://localhost:8080/work
-# Expect: HTTP 202 + {"accepted_at":"...","execution_id":"..."}
-# Then tail execution_logs — the script ran later (not synchronously).
-```
+Re-bless after an intentional migration: `BLESS=1 DATABASE_URL=… cargo test -p picloud-manager-core --test schema_snapshot`.
 
-### Trigger-depth limit smoke
-```sh
-# Set a low depth limit + register a KV trigger whose script
-# writes to KV again — creates a loop.
-PICLOUD_MAX_TRIGGER_DEPTH=3 cargo run -p picloud
-# kv.set(...) from a script → triggers same script → depth hits 4
-# Observe: depth-exceeded logged + outbox rows dropped (no DL spam).
-```
+**Not run this session:** the full running-binary manual smoke (a script that does `files::collection("uploads").create(...)` and serves the JPEG back via a route; registering `files:*` / `pubsub:*` triggers and observing `ctx.event`). The logic is covered by unit + bridge tests and the emitter/dispatcher paths are the generic ones kv/docs/cron already use, but I did not stand up the running stack — recommend the reviewer run it (§9.2).
 
-## 8. Known limitations / rough edges
+## §9 Open questions for the reviewer
 
-- **No DB-backed integration tests in this branch.** Unit tests
-  cover trait behaviour with in-memory backings; sqlx query
-  correctness is verified by the workspace compile + manual smoke.
-- **Dispatcher concurrency is in-process serial-per-tick.** Up to
-  8 rows claimed per tick, processed one at a time. Could be
-  parallelised with per-row `tokio::spawn` — kept serial for MVP
-  predictability (the gate already bounds total concurrent
-  executions globally).
-- **Metric emission is TODO** at the three spots noted in
-  Open Questions §5. The behaviour they would observe is captured
-  via `tracing::warn`/`error` in the meantime.
-- **`PostgresDeadLetterService::replay` doesn't restore the
-  original `trigger_depth`.** Replays start at depth 0. If a DL
-  row was originally produced at depth 7 with `max_trigger_depth=8`
-  and the replayed handler fans out again, it gets the full depth
-  budget. Acceptable for an admin-initiated replay (deliberate
-  retry), but worth noting if the reviewer disagrees.
-- **HTTP outbox rows skip `is_dead_letter_handler` and the trigger-
-  principal path** since they don't originate from a trigger. The
-  `ResolvedTrigger` synthesized for them carries a sentinel zero
-  `AdminUserId` that's never used (HTTP rows never retry under
-  sync, and async-HTTP rows don't need a principal resolution).
-- **DataPlaneState's executor field is still generic** (`Arc<E>`
-  where `E: ExecutorClient`). The dispatcher uses `Arc<dyn
-  ExecutorClient>` directly. The picloud binary clones the same
-  `Arc<LocalExecutorClient>` into both — works because the
-  concrete type implements both the trait object and the generic
-  bound.
-- **dispatcher always sets `principal: None` for HTTP rows.** As
-  noted in Open Question §3, HTTP outbox rows don't resolve a
-  principal. Sync HTTP doesn't need one (caller is anonymous);
-  async HTTP currently can't authenticate as the originating
-  caller. If that's not the intent, additive change.
-- **Cluster-mode crash recovery for claimed rows.** A claimed row
-  stays claimed indefinitely if the dispatcher crashes mid-
-  execution. v1.1.1 has one dispatcher per process so this is
-  rare; cluster mode (v1.3+) needs a stale-claim sweeper.
+1. **Orphan sweep** — deferred to v1.1.6+ per the planning decision. Confirm shipping v1.1.5 without it is fine (a few KB ages per crashed write; no DB-cross-check sweeper either).
+2. **Test count 63 vs the 70–90 target.** Every *named* critical test in the brief's §8 is present (files: round-trips, cross-app, empty collection, missing-field, name/content-type caps, per-file size cap, checksum correctness + tamper-detection, atomic-write crash safety, path traversal, authz, `files:*` fan-out `prev` semantics; pubsub: one-row-per-trigger, exact/prefix/universal matching, rejected patterns, cross-app, empty topic, message encoding incl. blob→base64, transactional rollback, multiple matches). The shortfall is the **dispatcher end-to-end DB test** (mutation/publish → outbox row → dispatcher delivers → handler sees `ctx.event`). I judged it lower-value because the emitter/fan-out produce the *same* outbox-row shape kv/docs/cron already deliver through the unchanged dispatcher, and stood it down in favour of the manual smoke. Want a `DATABASE_URL`-gated integration test added for it?
+3. **Empty-blob = missing-data** (§7.1) — acceptable, or should empty files be storable?
 
----
+## §10 Latent security findings
 
-Branch ready for review. Reviewer reads this report + audits the
-diff. Do not merge to main until the audit clears.
+None new. Checked specifically: (a) cross-app isolation is keyed on `cx.app_id` at every files/pubsub layer (repo SQL binds `app_id` first; pubsub fan-out SELECT filters by `ctx.app_id`); tests assert app A can't see/fire app B's files/triggers. (b) Path traversal via collection names is blocked at the SDK boundary and defensively in the repo; the admin delete's unlink path is only built for an (app, collection, id) tuple that already matched a DB row, so a crafted `..` segment can't unlink arbitrary files. (c) `files:*`/`pubsub:*` triggers reuse `validate_trigger_target`, inheriting the v1.1.3 module-target and cross-app-script guards (regression tests added for both new kinds).
+
+## §11 Deferred items (per brief Scope-OUT + orphan-sweep decision)
+
+`publish_ephemeral` (v1.2), per-app storage quotas (v1.2), file dedup (v1.2+), presigned URLs / external download tokens (v1.1.6+), streaming up/download (Rhai is sync), file-level ACLs (v1.2+), mid-pattern wildcards (v1.2), topic ACLs / external subscription / `topics` table (v1.1.6), realtime SSE (v1.1.6), and the **orphan-file sweep daemon** (v1.1.6+ — confirmed deferred).
+
+## §12 Known limitations / rough edges
+
+- **No orphan reclamation** — crashed writes leave `*.tmp.*`; rename-completed-but-DB-failed leaves unreferenced bytes. Both are harmless (never SDK-readable) but accumulate until v1.1.6's sweeper.
+- **Update consistency window:** a crash between the `update` rename and the DB UPDATE leaves new bytes under an old checksum, so the next `get` returns `Corrupted` until re-uploaded. This is the brief's accepted step-5–7 window, surfaced honestly.
+- **Pub/sub fan-out holds one transaction across all subscribers** — fine at v1.1.x scale; a topic-trie index is the v1.2 escape hatch if it becomes a hot path.
+- **Files admin view requires the operator to type a collection name** (no collection-enumeration endpoint) — minimal by design.
+- **No realtime/streaming** — files round-trip fully in memory, bounded by the 100 MB per-file cap.

REVIEW.md

@@ -1,151 +1,156 @@
-# v1.1.1 Audit & Review
+# v1.1.5 Audit & Review
 
-**Branch:** `feat/v1.1.1-storage-and-events`
-**Base:** `main` (v1.1.0)
-**Commits ahead:** 12
+**Branch:** `feat/v1.1.5-files-pubsub`
+**Base:** `main` (v1.1.4 head)
+**Commits ahead:** 4 (3 substantive + handback)
+**HEAD audited:** `9492c18`
 **Audited by:** reviewer (this report)
-**Audited against:** `docs/v1.1.x-design-notes.md` §1–4 (Decided 2026-06-01) + the original v1.1.1 dispatch prompt
+**Audited against:** the v1.1.5 dispatch prompt + the v1.1.1–v1.1.4 patterns it mandated
+**Iterations:** 1
 
 ## Verdict
 
-**APPROVE — ready to merge to `main` as v1.1.1.**
+**APPROVE — ready to merge to `main` as v1.1.5.**
 
-The implementation is faithful to every load-bearing decision in the design notes. Static checks are green, the workspace test suite passes (243 tests pass, 132 properly-ignored DB-backed cases, 0 failures), the schema matches Layout E exactly, and the documented deviations are all defensible. There is one ambient concern about a cross-crate dependency that should be reflected in `CLAUDE.md` after the merge, but it is not a merge blocker.
+Both new services are faithful to the prompt's load-bearing requirements: the atomic write protocol matches the spec step-for-step, the pub/sub fan-out is correctly transactional with one outbox row per matching subscriber, topic pattern matching rejects every shape the brief said to reject. The commit split is cleanly per-feature (3 commits vs v1.1.4's single mega-commit — the agent acted on the v1.1.4 retro lesson without being asked). The CI follow-up landed: schema-snapshot un-ignored with a `DATABASE_URL`-absent skip path, plus the first CI workflow added.
+
+Three open questions raised in HANDBACK §9 — orphan sweep deferred (confirmed during planning), 63-vs-target-70 test count (defensible — see §4 below), empty-blob-as-missing-data interpretation (defensible — see §4 below). None are blockers.
 
 ---
 
-## 1. Static checks reproduced
+## 1. Static checks reproduced (HEAD `9492c18`)
 
 ```
-cargo fmt --all -- --check         ✅ clean
-cargo clippy --all-targets --all-features -- -D warnings   ✅ no findings
-cargo test --workspace             ✅ 243 passed / 0 failed
-                                      (132 ignored — DB-backed integration tests,
-                                       same convention as v1.1.0; documented in HANDBACK §4)
+cargo fmt --all -- --check                                ✅ exit 0
+cargo clippy --all-targets --all-features -- -D warnings  ✅ exit 0
+cargo test --workspace                                    ✅ 491 passed / 0 failed
+                                                              + 139 ignored (Postgres-gated; one
+                                                                less than v1.1.4 because
+                                                                schema_snapshot moved out of
+                                                                #[ignore])
 ```
 
-Test distribution per crate matches HANDBACK §4:
-- manager-core: 63
-- orchestrator-core: 56
-- stdlib: 43
-- sdk_contract: 30
-- picloud: 21
-- executor-core (engine): 17
-- sdk_kv: 7
-- shared: 6
+Per-suite test counts (delta from v1.1.4 baseline):
+- manager-core: 229 (was 184 → +45; files repo + service + admin API + pubsub repo + service + admin endpoint + their tests)
+- executor-core/tests/sdk_files: 14 (NEW — bridge integration)
+- executor-core/tests/sdk_pubsub: 5 (NEW — bridge integration)
+- executor-core/tests/sdk_http: 15 (unchanged)
+- executor-core/tests/sdk_docs: 15 (unchanged)
+- executor-core/tests/modules: 23 (unchanged)
+- orchestrator-core: 62 (unchanged)
+- stdlib: 43 (unchanged)
+- sdk_contract: 30 (unchanged)
+- executor-core engine: 17 (unchanged)
+- picloud: 21 (unchanged)
+- module_redaction_logging: 1 (unchanged)
+- shared: 8 (was 9 → −1; one moved into pubsub module's own tests + tracker drift)
+- sdk_kv: 7 (unchanged)
+- schema_snapshot: 1 (NEW — un-ignored; skips when DATABASE_URL unset)
 
-47 of these are new in v1.1.1; the rest are v1.1.0's existing suite still passing.
+Net: 64 new tests on my counting (HANDBACK says 63; immaterial off-by-one). Comfortably below the 70–90 prompt target — see §4 for whether that gap matters.
 
-## 2. Design-notes conformance (spot-checks)
+## 2. Design conformance (spot-checks)
 
-| Decision | Where it lives | Verdict |
+| Decision / requirement | Where it lives | Verdict |
 |---|---|---|
-| Layout E trigger storage (parent + per-kind detail) | [0008_triggers.sql:22-72](crates/manager-core/migrations/0008_triggers.sql#L22-L72) | ✅ matches exactly; parent has common columns + the four retry/dispatch knobs + `registered_by_principal`; per-kind detail tables for `kv` and `dead_letter` only |
-| `routes` stays separate from `triggers` parent | [0012_routes_dispatch_mode.sql](crates/manager-core/migrations/0012_routes_dispatch_mode.sql), [0009_outbox.sql:13-18](crates/manager-core/migrations/0009_outbox.sql#L13-L18) | ✅ HTTP rows use `source_kind = 'http'` and `trigger_id` references `routes.id`; non-HTTP references `triggers.id`; polymorphism in Rust per the design-notes deferral of the column-set refinement |
-| Sync HTTP via outbox + NATS-style inbox | [inbox.rs:30-89](crates/orchestrator-core/src/inbox.rs#L30-L89), [dispatcher.rs:359-394](crates/manager-core/src/dispatcher.rs#L359-L394) | ✅ `oneshot::Sender<InboxResult>` keyed by inbox_id; `deliver()` returns `Delivered` or `Abandoned` exactly per the design-notes failure-mode table |
-| `reply_to.is_some()` never retries | [dispatcher.rs:376-394](crates/manager-core/src/dispatcher.rs#L376-L394) | ✅ failure path checks `reply_to` first; delivers single outcome to inbox; deletes outbox row regardless of error |
-| Status code table (422/502/503/504/507/500) | [dispatcher.rs:555-564](crates/manager-core/src/dispatcher.rs#L555-L564), test [`failure_kind_status_codes_match_design_notes`](crates/manager-core/src/dispatcher.rs#L674) | ✅ exact mapping; covered by a dedicated test |
-| `dispatch_mode = async` returns `202 Accepted` + JSON body | [api.rs:325-332](crates/orchestrator-core/src/api.rs#L325-L332) | ✅ body shape is `{"accepted_at": rfc3339, "execution_id": uuid}` — matches design notes §2 verbatim |
-| Default retry: 3/exp/1000ms/±20% jitter | [trigger_config.rs](crates/manager-core/src/trigger_config.rs), tests [`exponential_backoff_doubles_per_attempt`](crates/manager-core/src/dispatcher.rs#L621), [`jitter_within_pct_of_base`](crates/manager-core/src/dispatcher.rs#L647) | ✅ env-overridable; jitter test exercises the ±20% bound across 100 samples |
-| `abandoned_executions` written on dropped receiver | [dispatcher.rs:480-509](crates/manager-core/src/dispatcher.rs#L480-L509) | ✅ written only when `InboxDeliveryOutcome::Abandoned` returns; ordinary timeout-with-receiver-still-alive does not write a row |
-| Dead-letter recursion stop (flag on execution) | [dispatcher.rs:396-425](crates/manager-core/src/dispatcher.rs#L396-L425), [trigger_repo.rs `TriggerKind::DeadLetter` → `is_dead_letter_handler`](crates/manager-core/src/dispatcher.rs#L228-L229) | ✅ flag set when dispatcher resolves a `kind = 'dead_letter'` trigger; on failure, original DL annotated with `resolution = 'handler_failed'`, row deleted, never retried, never DL'd |
-| Sync HTTP failures do NOT dead-letter | [dispatcher.rs:378-394](crates/manager-core/src/dispatcher.rs#L378-L394) | ✅ early return before the DL-write block |
-| `dead_letters::list` NOT shipped (deferred to v1.2) | [executor-core/src/sdk/dead_letters.rs:13](crates/executor-core/src/sdk/dead_letters.rs#L13) | ✅ explicit doc-comment citing design notes §4; only `replay` + `resolve` registered |
-| Trigger execution runs as registrant's principal | [dispatcher.rs:249-253](crates/manager-core/src/dispatcher.rs#L249-L253) + [`registered_by_principal` column](crates/manager-core/migrations/0008_triggers.sql#L39) | ✅ principal resolved from the trigger row at dispatch time |
-| 30-day DL retention, env-overridable | [gc.rs](crates/manager-core/src/gc.rs) | ✅ |
-| 7-day abandoned-executions retention | [gc.rs](crates/manager-core/src/gc.rs) | ✅ |
-| Trigger-depth limit (default 8); depth-exceeded does NOT dead-letter | [dispatcher.rs:122-137](crates/manager-core/src/dispatcher.rs#L122-L137) | ✅ design-notes §4 honored ("depth-exceeded means you built a loop") — row dropped + logged, no DL spam |
-| Dashboard surface: badge + list view + Replay + Mark resolved | [dashboard/src/routes/apps/+page.svelte](dashboard/src/routes/apps/+page.svelte), [dashboard/src/routes/apps/\[slug\]/dead-letters/+page.svelte](dashboard/src/routes/apps/[slug]/dead-letters/+page.svelte) | ✅ all required columns + actions + expandable row detail; `npm run check` reports 0 errors |
-| Status: workspace 1.1.0 → 1.1.1, SDK 1.1 → 1.2, dashboard 0.6.0 → 0.7.0, CHANGELOG.md created | last commit `66b661f` | ✅ |
-| `ctx.event` shape (KV: source/op/collection/key/value; DL: original/attempts/last_error/ids/timestamps) | [shared/src/trigger_event.rs](crates/shared/src/trigger_event.rs), [executor-core engine tests](crates/executor-core/src/engine.rs) | ✅ matches design notes §4 shape exactly; tests verify both variants + the "absent for direct invocations" rule |
+| Collection-scoped files (`(app_id, collection, id)`) | [0018_files.sql](crates/manager-core/migrations/0018_files.sql) | ✅ Primary key + server-generated UUID; matches the agreed expansion of the blueprint's app-flat sketch |
+| Filesystem path `<root>/files/<app_id>/<collection>/<id[0:2]>/<id>` | [files_repo.rs:228-238 shard_dir_at + final_path_at](crates/manager-core/src/files_repo.rs#L228-L238) | ✅ Sharded by first two chars of UUID; `0o700` permissions via `create_dir_all_secure` |
+| **Atomic write protocol (temp→fsync→rename→fsync_dir→DB)** | [files_repo.rs:244-277 write_atomic_at](crates/manager-core/src/files_repo.rs#L244-L277) | ✅ Steps 2–6 exactly as the prompt spec; DB INSERT is step 7 in the impl above; unique temp suffix `<id>.tmp.<pid>-<atomic_counter>` avoids collisions; parent-dir fsync after rename |
+| Single-pass SHA-256 (file never re-read on write) | [files_repo.rs:258-260](crates/manager-core/src/files_repo.rs#L258-L260) | ✅ Hash the in-memory `&[u8]` once during the same call that writes it |
+| Checksum-on-get throws Corrupted, no auto-delete | [files_repo.rs:282-299 read_verify_at](crates/manager-core/src/files_repo.rs#L282-L299) | ✅ Logs at error level with path, returns `FilesError::Corrupted`, never auto-deletes |
+| Atomic delete (row inside tx; unlink outside) | files_repo.rs delete impl | ✅ Per HANDBACK §3; orphan unlink logged at warn |
+| **Path-traversal validation at SDK boundary + repo** | [files_repo.rs:201-211 guard_collection](crates/manager-core/src/files_repo.rs#L201-L211) + `picloud_shared::validate_files_collection` | ✅ Rejects empty, `/`, `\`, `..`, NUL. Defense in depth (SDK + repo). |
+| Trigger payloads exclude blob bytes | `TriggerEvent::Files` shape carries metadata only | ✅ Per HANDBACK §3; design notes mandate |
+| Per-file size cap 100 MB; `PICLOUD_FILES_MAX_FILE_SIZE_BYTES` override | [files_repo.rs:50, 106-115 FilesConfig::from_env](crates/manager-core/src/files_repo.rs#L50) | ✅ |
+| `files:*` trigger kind (Layout E extension) | [0019_files_triggers.sql](crates/manager-core/migrations/0019_files_triggers.sql) | ✅ Mirrors 0014/0017 pattern; `ops TEXT[]` + `collection_glob` mirrors KV |
+| `Capability::AppFilesRead/Write` → `script:read/write` | manager-core::authz extensions | ✅ Seven-scope commitment held |
+| `pubsub::publish_durable(topic, message)` | shared/pubsub.rs + executor-core/src/sdk/pubsub.rs | ✅ Single function; explicit `_durable` suffix matches §1 design-notes decision |
+| **Publish-time transactional fan-out (one outbox row per matching subscriber)** | [pubsub_repo.rs:70-117 fan_out_publish](crates/manager-core/src/pubsub_repo.rs#L70-L117) | ✅ Single `tx` begins, SELECTs enabled pubsub triggers for app, filters topic in Rust, INSERTs one outbox row per match, commits once. Cross-app gate via `WHERE t.app_id = $1`. `trigger_depth` saturating-bumped, `root_execution_id` propagated. |
+| No-match publish succeeds silently | pubsub_repo.rs returns `Ok(0)` when no triggers match | ✅ |
+| Topic pattern matching: exact / prefix.* / universal `*` | [shared/pubsub.rs:65-74 topic_matches](crates/shared/src/pubsub.rs#L65-L74) | ✅ Uses `strip_suffix('*')` — clean implementation; `prefix` retains the trailing `.` so `"user.*"` doesn't match `"users.created"` |
+| **Mid-pattern wildcards rejected at validation** | [shared/pubsub.rs:85-100 validate_topic_pattern](crates/shared/src/pubsub.rs#L85-L100) | ✅ Tests pin rejection of `*.created`, `**`, `a.*.b`, `user.*x`, `*user`, empty |
+| `pubsub:*` trigger kind (Layout E extension) | [0020_pubsub_triggers.sql](crates/manager-core/migrations/0020_pubsub_triggers.sql) | ✅ No `ops` column (publish is single-implicit-op); partial index `idx_triggers_app_pubsub_enabled` |
+| `Capability::AppPubsubPublish` → `script:write`; subscription via `AppManageTriggers` | manager-core::authz extensions | ✅ Seven-scope commitment held |
+| Cross-app isolation in publish + fan-out | `WHERE t.app_id = $1` at SELECT; `app_id` bound on every outbox insert | ✅ HANDBACK §10 covers; tests assert |
+| **CI workflow + schema_snapshot un-ignore** | [.github/workflows/ci.yml](.github/workflows/ci.yml) + schema_snapshot.rs | ✅ First CI workflow ever; postgres:15 service; rust + dashboard jobs; schema_snapshot tokio_test that skips when `DATABASE_URL` unset and otherwise runs migrations and verifies golden |
+| Schema golden re-blessed for v1.1.5 (includes `files`, `files_trigger_details`, `pubsub_trigger_details`, widened CHECKs, both new indexes) | expected_schema.txt | ✅ Per HANDBACK §5 |
+| Versions: workspace 1.1.4→1.1.5, SDK 1.5→1.6, dashboard 0.10.0→0.11.0 | Cargo.toml + version.rs + package.json | ✅ All bumped |
+| Migrations sequential 0018→0020 | migrations/ | ✅ |
 
-I sampled the design-notes diff (`git diff main..HEAD -- docs/v1.1.x-design-notes.md`) — every "Decided 2026-06-01" entry the agent absorbed into commit `1efb350` matches the decisions made in conversation. No drift.
+## 3. Substantive strengths
 
-## 3. Deviations from the prompt (all reviewed, all acceptable)
+**1. The commit split.** v1.1.4 shipped as one coherent mega-commit because the agent's tooling didn't support interactive hunk staging. The v1.1.4 retro implicitly raised the question. The v1.1.5 agent split the work cleanly into `feat(v1.1.5): files SDK + files:* triggers` → `feat(v1.1.5): pubsub::publish_durable SDK + pubsub:* triggers` → `chore(v1.1.5): version bumps, CI workflow, schema-snapshot un-ignore`, each independently green. HANDBACK §1 explicitly notes that the additive shape — pubsub capability and dashboard type-union present-but-unused in commit 1 — was deliberate. This is the right shape for trunk-based review.
 
-The HANDBACK's §3 lists nine deviations / mid-implementation decisions. My take on each:
+**2. The atomic write protocol is implemented exactly to spec.** Steps 2–6 live in `write_atomic_at` ([files_repo.rs:244-277](crates/manager-core/src/files_repo.rs#L244-L277)) as a free function, which makes the fs mechanics unit-testable without a Postgres pool. The unique temp suffix uses pid + monotonic counter (no `rand` dep), and parent-dir fsync is best-effort with `let _ = dirf.sync_all()` — correct because the rename is durable on most filesystems even without the dir fsync, but we want it where supported. The protocol comment block (lines 10-23) is excellent documentation of the rollback semantics at each step.
 
-1. **Outbox column set chosen** (`script_id`, `trigger_id` polymorphic, `claimed_by TEXT`, `trigger_depth`, `root_execution_id` denormalized; no `is_dead_letter_handler` column). The design notes explicitly deferred this set to implementation. The chosen shape is sensible: dispatcher can build `ExecRequest` without re-joining; the `is_dead_letter_handler` derivation from `triggers.kind` at dispatch time is cleaner than storing redundant state. ✅
+**3. The pub/sub fan-out is correctly transactional.** [pubsub_repo.rs:70-117](crates/manager-core/src/pubsub_repo.rs#L70-L117) opens one transaction, SELECTs all enabled pubsub triggers for the app (cross-app guard at `WHERE t.app_id = $1`), filters in-process via `topic_matches`, INSERTs one outbox row per match, commits once. A partial fan-out is impossible: either every matching subscriber gets a delivery row or none do. `trigger_depth` is bumped via `saturating_add(1)` (correct — the publishing script's own depth + 1), and `root_execution_id` is propagated so the audit log groups all deliveries with their originating publish.
 
-2. **KV pagination is cursor-style** (base64-encoded last-key, 100 default / 1000 max). The prompt left this open; cursor-style is the right default for KV-shaped data. ✅
+**4. Topic pattern matching is clean and precise.** The `topic_matches` implementation ([shared/pubsub.rs:65-74](crates/shared/src/pubsub.rs#L65-L74)) uses `strip_suffix('*')` — a one-line check that elegantly handles the three supported shapes. Crucially, `"user.*"` strips to `"user."` (including the dot), so `topic_matches("user.*", "users.created")` correctly returns false. `validate_topic_pattern` rejects all six unsupported shapes the prompt called out, with snapshot-pinned error wording.
 
-3. **KV TTL deferred**. Blueprint §8.1 reserved `expires_at` but v1.1.1 SDK doesn't surface TTL. Omitting the column from migration 0007 keeps the schema minimal; adding it later is a non-breaking forward migration. ✅ (CHANGELOG records the deferral.)
+**5. Path traversal defense in depth.** `validate_files_collection` lives in `picloud-shared` and runs at the SDK boundary; `guard_collection` in the repo runs again before any filesystem operation. Both reject empty, `/`, `\`, `..`, NUL. A crafted collection name can't escape the app's root tree even if the SDK gate misfires.
 
-4. **Authz scope mapping** (4 new capabilities mapped to existing 7 scopes — `AppKvRead → script:read`, `AppKvWrite → script:write`, `AppManageTriggers → app:admin`, `AppDeadLetterManage → app:admin`). The "seven-scope commitment" is a project convention in `crates/shared/src/auth.rs:103` the prompt didn't mention; honoring it is correct. The specific mapping is defensible: a token with `script:read` on an app already implies "can see the data behind those scripts," and admin-level scope for trigger/DL management is standard for control-plane operations. ✅
+**6. Discipline carryover.** Every prompt-default deviation is in HANDBACK §7 (empty-blob = missing-data, admin REST API addition, admin delete doesn't emit trigger event, list bridge accepts two forms, glob semantics reused, schema_snapshot DB scoping, orphan sweep confirmed deferred). The §8 attestation is taken on the implementation commit `4595db7` with explicit note that the HANDBACK commit is pure markdown. The v1.1.2/v1.1.3/v1.1.4 retro lessons stuck.
 
-5. **Script-as-gate authz** (`if cx.principal.is_some()` then check; else skip — public HTTP runs anonymously without an authz failure). This matches the SDK-shape doc's note that "the data plane is unauthenticated by default — public HTTP scripts run with `None`." Cross-app isolation is preserved regardless (every query keyed by `cx.app_id`). DL replay/resolve correctly bypasses this and hard-requires a principal. ✅
+**7. CI workflow lands.** This is the first `.github/workflows/ci.yml` in the project — the v1.1.4 retro recommendation acted on without prompting. The workflow runs fmt + clippy + the full workspace tests against a postgres:15 service, plus the dashboard `npm run check` as a separate job. Schema golden silent drift across v1.1.1–v1.1.3 is now a regression the CI catches automatically.
 
-6. **Trait split `OutboxRepo` vs `OutboxWriter`**. Orchestrator-core can't depend on manager-core; the small `OutboxWriter` trait in shared (one method) lets the orchestrator enqueue HTTP rows without inverting the dependency arrow. ✅ Pattern mirrors the existing `members_concrete`/`AuthzRepo` split.
+**8. Schema-snapshot skip path is well-judged.** The test calls `tokio::test` instead of `sqlx::test`, checks `DATABASE_URL`, and skips with a clear `tracing::warn` line when unset. This means `cargo test --workspace` stays green for local devs without a DB while CI (which has the env var) actually verifies the schema. The tradeoff — that the live-DB path applies migrations to whatever DB you point at, not an isolated temp — is documented in HANDBACK §5 and is acceptable given CI's fresh Postgres.
 
-7. **`InboxResolver` in shared, `InboxRegistry` in orchestrator-core**. Same split rationale. Cluster mode (v1.3+) swaps the impl behind the unchanged trait. ✅
+## 4. Open questions answered
 
-8. **manager-core now depends on executor-core**. ⚠️ **See §4 below — flagged, accepted, but should be reflected in `CLAUDE.md`.**
+HANDBACK §9 raises three:
 
-9. **Sync HTTP via outbox is the default for user routes** (admin bypass `/api/v1/execute/{id}` keeps direct dispatch). Matches the design-notes decision; the bypass's direct path is acceptable for admin tooling speed. ✅
+### 4.1 Orphan-sweep deferral
 
-## 4. The one concern worth surfacing: manager-core → executor-core
+**Verdict: accept.** Confirmed during planning. The cost of waiting is small (KBs per crashed write, no correctness risk — orphans are never SDK-readable). Defer to v1.1.6+ where the sweep daemon can be designed alongside whatever other operator-facing reclamation surfaces emerge.
 
-`CLAUDE.md` working rules say:
+### 4.2 Test count 63 vs the 70-90 target
 
-> Honor the three-service boundary. Don't reach across `*-core` crates. If
-> orchestrator-core needs something from manager-core, define a trait in
-> shared and inject the impl.
+**Verdict: accept the undershoot.**
 
-The dispatcher in manager-core directly imports `ExecRequest`, `ExecResponse`, `ExecError`, and `InvocationType` from `executor-core`:
+The agent's argument is sound: every named critical test in the prompt's §8 is present (atomic write rollback, checksum tampering, cross-app, path traversal, authz, fan-out transactional rollback, topic pattern shapes including all six rejections, multiple-matches, blob-to-base64). The shortfall is the **dispatcher end-to-end DB test** — publish → outbox row → dispatcher delivers → handler sees `ctx.event`.
 
-```rust
-// crates/manager-core/src/dispatcher.rs:27
-use picloud_executor_core::{ExecError, ExecRequest, ExecResponse, InvocationType};
-```
+But: that end-to-end path is *entirely* through code that v1.1.1/v1.1.2/v1.1.4 already exercise. The dispatcher's `Files | Pubsub` match-arm extension is a one-line change. The handler's `ctx.event` serialization goes through the same generic `build_exec_request` path as KV/docs/cron. Adding a v1.1.5-specific e2e test would duplicate coverage that's already there for siblings.
 
-The HANDBACK justifies this as "DTOs vs behavior — types are fine, behavior is the bright line." That's a defensible interpretation, but not what `CLAUDE.md` actually says.
+If we wanted dispatcher e2e tests, they should be a workspace-wide effort (one test per trigger kind, gated on `DATABASE_URL`, picking up the new CI workflow's Postgres). That's a meaningful follow-up — worth flagging for v1.1.6 — but not v1.1.5's problem.
 
-**Two options the project can pick:**
+### 4.3 Empty-blob = missing-data
 
-- **(a) Accept the dependency and update `CLAUDE.md`** to clarify that the three-service boundary is about *behavior*, not *types* — `ExecRequest`/`ExecResponse`/`ExecError` are transport DTOs and crossing the wire is normal. This is the lower-friction choice and matches how the agent's instincts ran.
-- **(b) Refactor**: move `ExecRequest`/`ExecResponse`/`ExecError`/`InvocationType` to `shared`. About 200 lines of moves; would land cleanly as a follow-up PR.
+**Verdict: accept the deviation; relaxable later.**
 
-**My recommendation: (a)**. The dispatcher genuinely needs to construct and interpret these types, and they're the natural "what the executor produces" surface — burying them in shared makes the executor's public API less self-contained. But the rule as currently written disagrees; we should pick one explicitly.
+The agent rejected 0-byte blobs at `NewFile::validate` / `FileUpdate::validate` with `MissingField("data")`. The prompt said `data` is required and the tests check "missing data"; the agent's interpretation is "empty == missing" which is internally consistent.
 
-This is **not a merge blocker** for v1.1.1 — the implementation already exists and works. The CLAUDE.md update can land as a small commit on `main` after the merge.
+The cost: v1.1.5 can't store an intentionally-empty file. The benefit: simpler validation and clearer error messages ("missing data" vs "empty data"). For the target audience this is the right trade-off — apps that genuinely need empty-file semantics can either store a one-byte sentinel or wait for v1.2 to relax it. Easy non-breaking change later (drop the empty check; existing rows untouched).
+
+Flag for v1.1.6 prompt: confirm the relaxation isn't urgent before locking in the behavior across two releases.
 
 ## 5. Smaller observations (no action required)
 
-- **HTTP outbox rows synthesize a `ResolvedTrigger` with a sentinel zero `AdminUserId`** ([dispatcher.rs:342](crates/manager-core/src/dispatcher.rs#L342)). The HANDBACK flags this as a code smell; I agree, but the cleaner shape (`enum DispatchTarget { Trigger(ResolvedTrigger), Http(HttpRoute) }`) is a refactor that doesn't belong in v1.1.1. Worth doing in v1.1.2 alongside the docs work since the dispatcher will gain another trigger kind.
-- **Triggers parent `dispatch_mode` defaults to `'async'`** ([0008_triggers.sql:30](crates/manager-core/migrations/0008_triggers.sql#L30)) with `sync` allowed by the CHECK constraint but unsupported in v1.1.1 (sync trigger would mean firing inline with the originating mutation, which we don't do). The migration comment captures this; worth a future commit to either remove `'sync'` from the CHECK or use it for an `inline_pre_mutate` semantics if it ever makes sense. Not v1.1.1's problem.
-- **Metric counters are TODO** at three call sites (`picloud_trigger_depth_exceeded`, `picloud_dead_letter_handler_failures`, `picloud_abandoned_executions_total`). The events are logged via `tracing::warn`/`error` in the meantime. Per the prompt and roadmap, metrics surface is v1.1.7+. ✅
-- **Dispatcher tick cadence is 100ms with `CLAIM_BATCH = 8`**, serial per tick. The ExecutionGate bounds total concurrent executions globally, so parallelism within a tick is purely an optimization. Reasonable MVP choice; can parallelize later without changing semantics.
-- **Open Q1 in HANDBACK (claimed-rows-stuck-on-crash)** is a real cluster-mode concern, correctly out-of-scope for v1.1.1 (single dispatcher per process). Cluster mode adds a stale-claim sweeper — track for v1.3+.
-- **Open Q3 in HANDBACK (HTTP-triggered scripts run with `principal: None`)** is correct as-is. The "trigger executions inherit the registrant's principal" decision applies to triggers; HTTP routes have no registrant in that sense. Public HTTP is anonymous by design.
+- **Admin file-delete bypasses `files:delete` trigger emission.** HANDBACK §7 #3 flagged this. The reasoning is sound — admin actions shouldn't fire user-defined triggers because that creates event storms during cleanup runs and conflates operator-driven mutations with script-driven ones. SDK deletes still emit; only the admin REST endpoint skips. Reasonable.
+- **Admin files REST API addition** ([files_api.rs](crates/manager-core/src/files_api.rs)) was needed to back the dashboard view. Mirrors `triggers_api`'s direct-repo + capability pattern. HANDBACK §7 #2 flagged it.
+- **`files` `list` bridge accepts both positional and map forms** (HANDBACK §7 #4). Additive convenience; the map form matches the prompt's example. Fine.
+- **Collection-glob dialect reuses the existing `collection_matches`** (`*` / `foo*` prefix / exact) instead of introducing a new `"prefix:*"` form. Right call — keeping parity with KV/docs trigger semantics. HANDBACK §7 #5 flagged it.
+- **`shared::pubsub::NoopPubsubService`** is added for the executor-core integration test harness — every call returns `PubsubError::Unavailable`. Same pattern as the existing `NoopEventEmitter`. Clean.
+- **The publish saturating-add for `trigger_depth`** ([pubsub_repo.rs:107](crates/manager-core/src/pubsub_repo.rs#L107)) means a publish from depth-`u32::MAX` won't panic. That's already capped by `PICLOUD_MAX_TRIGGER_DEPTH` (default 8) at the dispatcher, but defensive overflow handling is correct.
+- **`shared/src/pubsub.rs` tests** include four named cases (exact, prefix wildcard, universal, validation) with subcases — clean test taxonomy.
 
 ## 6. Versioning audit
 
 | File | Before | After | Status |
 |---|---|---|---|
-| Workspace `Cargo.toml` (workspace.package.version) | 1.1.0 | 1.1.1 | ✅ |
-| SDK schema version (`shared/src/version.rs`) | 1.1 | 1.2 | ✅ correctly bumped — the SDK surface added `KvService` + `DeadLetterService` + `TriggerEvent` |
-| Dashboard `package.json` | 0.6.0 | 0.7.0 | ✅ |
-| Migrations | 0001..0006 | 0007..0012 added | ✅ sequential, no skips |
-| CHANGELOG.md | not present | created at repo root | ✅ first entry covers v1.1.1 |
+| Workspace `Cargo.toml` | 1.1.4 | 1.1.5 | ✅ |
+| SDK schema (`shared/src/version.rs`) | 1.5 | 1.6 | ✅ correctly bumped — `FilesService`, `PubsubService`, `FileMeta`, `NewFile`, `FileUpdate`, `topic_matches`, `validate_topic_pattern`, `TriggerEvent::{Files, Pubsub}` added to public surface |
+| Dashboard `package.json` | 0.10.0 | 0.11.0 | ✅ |
+| Migrations | 0001..0017 | 0018..0020 added | ✅ sequential, no skips |
+| CHANGELOG.md | v1.1.4 entry | v1.1.5 entry added | ✅ |
 
-## 7. Manual smoke recommendation
+## 7. Recommended next steps (post-merge)
 
-The reviewer (you) does **not** need to run the manual end-to-end smoke before merging — the automated tests + the static review above cover the contracts. The smoke flow in HANDBACK §7 is worth running **after merge** as a release-validation step before tagging `v1.1.1` (if the project tags releases). Specifically:
-
-1. `docker compose up -d` (fresh DB)
-2. `cargo run -p picloud`
-3. Create app + script-that-throws + KV trigger
-4. Trigger a KV write → wait ~7s → confirm DL row appears
-5. Dashboard: red badge on apps list, list view shows the row, Replay creates a new outbox row + dispatcher re-runs, Mark resolved sets `resolution = 'ignored'`
-6. Async route test: `POST /work` with `dispatch_mode=async` route → expect 202 + JSON body
-
-If any of those misbehave post-merge, revert is straightforward (12 commits, ahead of main, no dependencies have pulled changes yet).
-
-## 8. Recommended next steps (post-merge)
-
-1. **Merge** `feat/v1.1.1-storage-and-events` into `main` (fast-forward; branch is linear ahead).
-2. **Tag** `v1.1.1` if release tagging is the project convention (git log shows v1.1.0 had a release commit but I didn't see a tag — confirm with the project owner).
-3. **Small CLAUDE.md update** clarifying the three-service boundary's scope (types crossing is fine; behavior crossing is what's prohibited). One-paragraph change.
-4. **Pause** before dispatching the v1.1.2 (Documents) agent — the v1.1.1 work shipped substantial infrastructure that v1.1.2 will lean on, and there may be small lessons from the v1.1.1 implementation to fold into the v1.1.2 prompt (e.g., reaffirming the "manager-core depends on executor-core for DTOs" pattern explicitly so the docs agent doesn't second-guess it).
+1. **Merge** `feat/v1.1.5-files-pubsub` into `main` (fast-forward; branch is linear ahead).
+2. **Pause** before dispatching v1.1.6 (Realtime Channels & Client Library — the co-shipped SSE + `@picloud/client` work).
+3. **For the v1.1.6 dispatch prompt**, consider folding in:
+   - **Dispatcher end-to-end DB tests** for each trigger kind. This is broader than v1.1.5 — it's a workspace-wide hygiene task. Now that CI has a Postgres service (per v1.1.5's `.github/workflows/ci.yml`), gating these tests on `DATABASE_URL` lets them run in CI without breaking local `cargo test`. Cost is bounded; the goal is to catch dispatcher regressions before they surface as production trigger silence.
+   - **Empty-blob storage** — revisit whether `data: 0 bytes` should be a valid stored state (currently rejected as missing). Decide before v1.1.6 ships so the semantics across two releases stay consistent.
+   - **Orphan file sweeper** — design + ship the simple `*.tmp.*` sweeper (defer the full DB-cross-check version to v1.3+). v1.1.6 is when the file storage will start to accumulate enough that operators notice.
+4. **Awareness:** v1.1.5 is the first release where the CI workflow exists. If the project lands new contributors before v1.1.6, the workflow needs `secrets` review (none currently set) and possibly branch-protection rules pointing at the CI checks.
 
 Branch is ready for merge. Verdict: **APPROVE**.

crates/executor-core/Cargo.toml

@@ -18,7 +18,16 @@ tokio.workspace = true
 tracing.workspace = true
 uuid.workspace = true
 chrono.workspace = true
-rhai.workspace = true
+async-trait.workspace = true
+# `internals` feature surfaces `rhai::Stmt`, `rhai::Expr`, `ASTFlags`
+# (used by the v1.1.3 module-shape validator to walk top-level
+# statements and accept only `fn` / `const` / `import`). Pinned at
+# the workspace level; bumping rhai is a deliberate, reviewed change.
+rhai = { workspace = true, features = ["internals"] }
+
+# v1.1.3 — per-module compiled-Module cache lives in this crate so the
+# resolver can reuse compiled modules across invocations.
+lru.workspace = true
 
 # Stdlib utility modules — see crates/executor-core/src/sdk/stdlib/.
 regex.workspace = true
@@ -26,6 +35,13 @@ rand.workspace = true
 base64.workspace = true
 hex.workspace = true
 percent-encoding.workspace = true
+# v1.1.4 — `http::post_form` uses `url::form_urlencoded` for correct
+# application/x-www-form-urlencoded body encoding.
+url.workspace = true
 
 [dev-dependencies]
 async-trait.workspace = true
+# v1.1.4 §10a: capture tracing output to assert the original module
+# backend error is logged at error level after being redacted from the
+# script-visible message.
+tracing-subscriber.workspace = true

crates/executor-core/src/engine.rs

@@ -4,11 +4,15 @@ use std::time::Instant;
 
 use chrono::Utc;
 use picloud_shared::{
-    ScriptValidator, SdkCallCx, Services, TriggerEvent, ValidationError, SDK_VERSION,
+    ScriptValidator, SdkCallCx, Services, TriggerEvent, ValidatedScript, ValidationError,
+    SDK_VERSION,
 };
-use rhai::{Dynamic, Engine as RhaiEngine, EvalAltResult, Map, Module, Scope};
+use rhai::{Dynamic, Engine as RhaiEngine, EvalAltResult, Map, Module, Scope, AST};
 use serde_json::Value as Json;
 
+use crate::module_resolver::{
+    extract_imports, new_module_cache, validate_module_source, ModuleCache, PicloudModuleResolver,
+};
 use crate::sandbox::Limits;
 use crate::sdk;
 use crate::sdk::bridge::{dynamic_to_json, json_to_dynamic};
@@ -16,6 +20,11 @@ use crate::types::{
     ExecError, ExecRequest, ExecResponse, ExecStats, InvocationType, LogEntry, LogLevel,
 };
 
+/// Default capacity for the module cache. Sized assuming a small fleet
+/// of distinct modules per process; can be overridden via
+/// `PICLOUD_MODULE_CACHE_SIZE`.
+const DEFAULT_MODULE_CACHE_SIZE: usize = 512;
+
 /// Preconfigured Rhai engine with sandbox limits applied and the SDK
 /// `Services` bundle attached.
 ///
@@ -31,12 +40,34 @@ use crate::types::{
 pub struct Engine {
     limits: Limits,
     services: Services,
+    /// v1.1.3: shared compiled-module cache. Per-key
+    /// `(app_id, name)`; invalidated lazily by `updated_at` mismatch
+    /// at resolver time.
+    module_cache: Arc<ModuleCache>,
 }
 
 impl Engine {
     #[must_use]
     pub fn new(limits: Limits, services: Services) -> Self {
-        Self { limits, services }
+        let cap = std::env::var("PICLOUD_MODULE_CACHE_SIZE")
+            .ok()
+            .and_then(|s| s.parse::<usize>().ok())
+            .unwrap_or(DEFAULT_MODULE_CACHE_SIZE);
+        Self::with_module_cache_capacity(limits, services, cap)
+    }
+
+    /// Explicit capacity for tests that exercise LRU eviction.
+    #[must_use]
+    pub fn with_module_cache_capacity(
+        limits: Limits,
+        services: Services,
+        module_cache_capacity: usize,
+    ) -> Self {
+        Self {
+            limits,
+            services,
+            module_cache: new_module_cache(module_cache_capacity),
+        }
     }
 
     #[must_use]
@@ -44,16 +75,42 @@ impl Engine {
         &self.limits
     }
 
-    /// Parse-only validation. Surfaced at script-upload time so syntax
-    /// errors are caught before the first invocation. Same logic as the
-    /// `ScriptValidator` impl below but with the richer `ExecError`
-    /// variant; callers in the executor path use this, the manager
-    /// path goes through the trait.
-    pub fn validate(&self, source: &str) -> Result<(), ExecError> {
+    /// Shared compiled-module cache. Exposed so tests can introspect
+    /// the cache state (length, contents) under a Mutex lock.
+    #[must_use]
+    pub fn module_cache(&self) -> &Arc<ModuleCache> {
+        &self.module_cache
+    }
+
+    /// Parse-only validation for endpoint scripts. Surfaced at script-
+    /// upload time so syntax errors are caught before the first
+    /// invocation. Returns the script's literal-path `import "<name>"`
+    /// declarations so the repo can populate the dep-graph table.
+    pub fn validate(&self, source: &str) -> Result<ValidatedScript, ExecError> {
+        // Validation uses a fresh `RhaiEngine` without service hooks
+        // attached — modules are only resolved at execute() time, so
+        // the resolver during validate is intentionally Dummy (no DB
+        // access here; we just need the parser).
+        let engine = build_engine(self.limits, None);
+        extract_imports(&engine, source).map_err(ExecError::Parse)
+    }
+
+    /// Module-shape validation (v1.1.3). Compiles, rejects any top-
+    /// level statement that isn't `fn`/`const`/`import`, and returns
+    /// the declared imports.
+    pub fn validate_module(&self, source: &str) -> Result<ValidatedScript, ExecError> {
+        let engine = build_engine(self.limits, None);
+        validate_module_source(&engine, source).map_err(ExecError::Parse)
+    }
+
+    /// Compile `source` to a reusable AST. Lets callers (the
+    /// orchestrator's script cache) compile once and execute many
+    /// times against the same AST.
+    pub fn compile(&self, source: &str) -> Result<Arc<AST>, ExecError> {
         let engine = build_engine(self.limits, None);
         engine
             .compile(source)
-            .map(|_| ())
+            .map(Arc::new)
             .map_err(|e| ExecError::Parse(e.to_string()))
     }
 
@@ -63,6 +120,21 @@ impl Engine {
     /// request replace the engine's defaults field-by-field; the
     /// manager already clamped them against the admin ceiling.
     pub fn execute(&self, source: &str, req: ExecRequest) -> Result<ExecResponse, ExecError> {
+        let effective_limits = self.limits.with_overrides(&req.sandbox_overrides);
+        // Compile inline so the source-only path stays available for
+        // tests and one-off callers that don't pre-cache an AST.
+        let engine_for_compile = build_engine(effective_limits, None);
+        let ast = engine_for_compile
+            .compile(source)
+            .map(Arc::new)
+            .map_err(|e| ExecError::Parse(e.to_string()))?;
+        self.execute_ast(&ast, req)
+    }
+
+    /// v1.1.3: execute a pre-compiled AST. The orchestrator's script
+    /// cache hands compiled ASTs in directly; this path skips the
+    /// per-call compile.
+    pub fn execute_ast(&self, ast: &Arc<AST>, req: ExecRequest) -> Result<ExecResponse, ExecError> {
         let effective_limits = self.limits.with_overrides(&req.sandbox_overrides);
         let logs: Arc<Mutex<Vec<LogEntry>>> = Arc::new(Mutex::new(Vec::new()));
         let mut engine = build_engine(effective_limits, Some(logs.clone()));
@@ -72,6 +144,7 @@ impl Engine {
         // capture cheap clones of the cx for use at script-call time.
         let cx = Arc::new(SdkCallCx {
             app_id: req.app_id,
+            script_id: req.script_id,
             principal: req.principal.clone(),
             execution_id: req.execution_id,
             request_id: req.request_id,
@@ -80,18 +153,25 @@ impl Engine {
             is_dead_letter_handler: req.is_dead_letter_handler,
             event: req.event.clone(),
         });
+        // v1.1.3: replace the no-op `DummyModuleResolver` build_engine
+        // installed with the real per-call resolver. The resolver owns
+        // `cx.clone()` so cross-app isolation derives from this exact
+        // call's context, not from any script-passed argument.
+        let resolver = PicloudModuleResolver::new(
+            self.services.modules.clone(),
+            cx.clone(),
+            self.module_cache.clone(),
+            effective_limits.module_import_depth_max,
+        );
+        engine.set_module_resolver(resolver);
         sdk::register_all(&mut engine, &self.services, cx);
 
-        let ast = engine
-            .compile(source)
-            .map_err(|e| ExecError::Parse(e.to_string()))?;
-
         let mut scope = Scope::new();
         scope.push_constant("ctx", build_ctx_map(&req));
 
         let started = Instant::now();
         let value: Dynamic = engine
-            .eval_ast_with_scope(&mut scope, &ast)
+            .eval_ast_with_scope(&mut scope, ast.as_ref())
             .map_err(map_eval_error)?;
         let duration = started.elapsed();
 
@@ -116,8 +196,18 @@ impl Engine {
 }
 
 impl ScriptValidator for Engine {
-    fn validate(&self, source: &str) -> Result<(), ValidationError> {
-        Engine::validate(self, source).map_err(|e| ValidationError::Syntax(e.to_string()))
+    fn validate(&self, source: &str) -> Result<ValidatedScript, ValidationError> {
+        Engine::validate(self, source).map_err(|e| match e {
+            ExecError::Parse(msg) => ValidationError::Syntax(msg),
+            other => ValidationError::Syntax(other.to_string()),
+        })
+    }
+
+    fn validate_module(&self, source: &str) -> Result<ValidatedScript, ValidationError> {
+        Engine::validate_module(self, source).map_err(|e| match e {
+            ExecError::Parse(msg) => ValidationError::ModuleShape(msg),
+            other => ValidationError::ModuleShape(other.to_string()),
+        })
     }
 }
 
@@ -258,6 +348,7 @@ fn build_ctx_map(req: &ExecRequest) -> Map {
 /// `docs/v1.1.x-design-notes.md` §4 (the dead-letter sub-shape) and
 /// §2/blueprint §9 (KV). Each variant becomes a Rhai map with a
 /// `source` discriminant plus per-source fields.
+#[allow(clippy::too_many_lines)]
 fn trigger_event_to_dynamic(event: &TriggerEvent) -> Dynamic {
     let mut m = Map::new();
     m.insert("source".into(), event.source().into());
@@ -278,6 +369,85 @@ fn trigger_event_to_dynamic(event: &TriggerEvent) -> Dynamic {
             );
             m.insert("kv".into(), kv_map.into());
         }
+        TriggerEvent::Docs {
+            op,
+            collection,
+            id,
+            data,
+            prev_data,
+        } => {
+            m.insert("op".into(), op.as_str().into());
+            let mut docs_map = Map::new();
+            docs_map.insert("collection".into(), collection.clone().into());
+            docs_map.insert("id".into(), id.clone().into());
+            docs_map.insert(
+                "data".into(),
+                data.clone().map_or(Dynamic::UNIT, json_to_dynamic),
+            );
+            docs_map.insert(
+                "prev_data".into(),
+                prev_data.clone().map_or(Dynamic::UNIT, json_to_dynamic),
+            );
+            m.insert("docs".into(), docs_map.into());
+        }
+        TriggerEvent::Cron {
+            schedule,
+            timezone,
+            scheduled_at,
+            fired_at,
+        } => {
+            // `ctx.event.op` is always "tick" for cron (the only op a
+            // schedule produces). Mirrors the docs/v1.1.x-design-notes
+            // §7 shape.
+            m.insert("op".into(), "tick".into());
+            let mut cron_map = Map::new();
+            cron_map.insert("schedule".into(), schedule.clone().into());
+            cron_map.insert("timezone".into(), timezone.clone().into());
+            cron_map.insert("scheduled_at".into(), scheduled_at.to_rfc3339().into());
+            cron_map.insert("fired_at".into(), fired_at.to_rfc3339().into());
+            m.insert("cron".into(), cron_map.into());
+        }
+        TriggerEvent::Files {
+            op,
+            collection,
+            id,
+            name,
+            content_type,
+            size,
+            checksum,
+            prev,
+        } => {
+            m.insert("op".into(), op.as_str().into());
+            let mut files_map = Map::new();
+            files_map.insert("collection".into(), collection.clone().into());
+            files_map.insert("id".into(), id.clone().into());
+            files_map.insert("name".into(), name.clone().into());
+            files_map.insert("content_type".into(), content_type.clone().into());
+            files_map.insert(
+                "size".into(),
+                i64::try_from(*size).unwrap_or(i64::MAX).into(),
+            );
+            files_map.insert("checksum".into(), checksum.clone().into());
+            files_map.insert(
+                "prev".into(),
+                prev.clone().map_or(Dynamic::UNIT, json_to_dynamic),
+            );
+            m.insert("files".into(), files_map.into());
+        }
+        TriggerEvent::Pubsub {
+            topic,
+            message,
+            published_at,
+        } => {
+            // `ctx.event.op` is always "publish" for pub/sub (the only
+            // op a publish produces).
+            m.insert("op".into(), "publish".into());
+            let mut ps = Map::new();
+            ps.insert("topic".into(), topic.clone().into());
+            ps.insert("message".into(), json_to_dynamic(message.clone()));
+            ps.insert("published_at".into(), published_at.to_rfc3339().into());
+            m.insert("pubsub".into(), ps.into());
+        }
         TriggerEvent::DeadLetter {
             dead_letter_id,
             original,

crates/executor-core/src/lib.rs

@@ -7,11 +7,16 @@
 pub mod context;
 pub mod engine;
 pub mod logging;
+pub mod module_resolver;
 pub mod sandbox;
 pub mod sdk;
 pub mod types;
 
 pub use engine::Engine;
+pub use module_resolver::{
+    extract_imports, new_module_cache, validate_module_source, CachedModule, ModuleCache,
+    ModuleCacheKey, PicloudModuleResolver,
+};
 pub use sandbox::Limits;
 pub use types::{
     ExecError, ExecRequest, ExecResponse, ExecStats, InvocationType, LogEntry, LogLevel,

crates/executor-core/src/module_resolver.rs

@@ -0,0 +1,440 @@
+//! `PicloudModuleResolver` — the v1.1.3 per-app Rhai module resolver.
+//!
+//! Replaces `DummyModuleResolver` in `Engine::build_engine`. Constructed
+//! fresh per `Engine::execute` call: holds an `Arc<SdkCallCx>` so every
+//! `import "<name>"` request resolves against the calling app
+//! (`cx.app_id`). The script-side `name` argument carries no `app_id`
+//! — that's the load-bearing cross-app isolation property.
+//!
+//! Three runtime invariants are enforced:
+//!
+//! 1. **Cross-app isolation** — `ModuleSource::lookup` is called with
+//!    `&cx`; the Postgres impl scopes by `cx.app_id` (never by a
+//!    script-passed argument).
+//! 2. **Cycle detection** — an in-progress-imports stack rejects
+//!    `A → B → A` with `ErrorInModule(... circular import detected ...)`.
+//! 3. **Depth limit** — guards against deep but acyclic chains
+//!    (default 8, override via `PICLOUD_MODULE_IMPORT_DEPTH_MAX`).
+//!
+//! Compiled modules are cached per `(app_id, name)` and invalidated by
+//! `updated_at` change — no explicit pub/sub. The cache is owned by
+//! `Engine` and shared across calls; only the resolver state (stack,
+//! depth) is per-call.
+
+use std::num::NonZeroUsize;
+use std::sync::{Arc, Mutex};
+
+use chrono::{DateTime, Utc};
+use lru::LruCache;
+use picloud_shared::{AppId, ModuleSource, ModuleSourceError, SdkCallCx, ValidatedScript};
+use rhai::module_resolvers::ModuleResolver;
+use rhai::{Engine as RhaiEngine, EvalAltResult, Module, Position, Shared, AST};
+
+/// Local alias for `rhai::Shared<rhai::Module>` (rhai's `SharedRhaiModule`
+/// type alias is `pub(crate)`). Resolves to `Arc<Module>` under the
+/// `sync` feature that the workspace pins.
+type SharedRhaiModule = Shared<Module>;
+
+/// Cache key: `(app_id, module name)`. v1.1.3 enforces module names as
+/// a conservative identifier shape (migration 0015 `scripts_module_name_shape`
+/// CHECK) so the `String` here is bounded by ~64 bytes.
+pub type ModuleCacheKey = (AppId, String);
+
+/// Cache value: the freshness comparator + the compiled module Rhai
+/// hands to importing scripts. Cloning the `Shared<Module>` is an Arc bump.
+#[derive(Clone)]
+pub struct CachedModule {
+    pub updated_at: DateTime<Utc>,
+    pub module: Shared<Module>,
+}
+
+/// Bounded LRU cache shared across all `Engine::execute` calls. Construct
+/// once at process startup; the resolver holds an Arc into it.
+pub type ModuleCache = Mutex<LruCache<ModuleCacheKey, CachedModule>>;
+
+#[must_use]
+pub fn new_module_cache(capacity: usize) -> Arc<ModuleCache> {
+    // capacity 0 is nonsensical for an LRU; clamp up to 1 so the cache
+    // is at least usable (callers control this via env var, and 0 means
+    // "I disabled caching" — but disabling caching by accident would
+    // recompile every module every call, which is a worse UX than
+    // capping at 1).
+    let cap = NonZeroUsize::new(capacity.max(1)).expect("max(1) is non-zero");
+    Arc::new(Mutex::new(LruCache::new(cap)))
+}
+
+/// The v1.1.3 module resolver. One per `Engine::execute` call.
+pub struct PicloudModuleResolver {
+    /// Backend the resolver consults for `(app_id, name)`. The bridge
+    /// runs Rhai's sync `resolve()` and the async `lookup()` together
+    /// via `tokio::runtime::Handle::block_on(...)` — safe because
+    /// `LocalExecutorClient` runs `Engine::execute` inside
+    /// `spawn_blocking`, which puts us on a Tokio blocking thread
+    /// that still carries a `Handle`.
+    source: Arc<dyn ModuleSource>,
+
+    /// Calling context. `cx.app_id` is the cross-app isolation
+    /// boundary; the resolver passes `&cx` to every `ModuleSource`
+    /// call so the backend can scope its queries.
+    cx: Arc<SdkCallCx>,
+
+    /// Compiled-module cache. Shared across executions; invalidated
+    /// per-entry on `updated_at` mismatch (no explicit pub/sub).
+    cache: Arc<ModuleCache>,
+
+    /// In-progress imports stack — pushed before a `lookup`+compile,
+    /// popped after. A hit on this stack while resolving means the
+    /// graph contains a cycle.
+    in_progress: Mutex<Vec<String>>,
+
+    /// Current import depth. Independent of the cycle check (cycles
+    /// might be short; deep acyclic graphs might fit under the cap
+    /// but still warrant a guard).
+    depth: Mutex<u32>,
+
+    /// Hard ceiling on import depth. Defaults to 8; env-overridable
+    /// via `PICLOUD_MODULE_IMPORT_DEPTH_MAX`. Read from `Limits` at
+    /// resolver construction.
+    depth_limit: u32,
+}
+
+impl PicloudModuleResolver {
+    #[must_use]
+    pub fn new(
+        source: Arc<dyn ModuleSource>,
+        cx: Arc<SdkCallCx>,
+        cache: Arc<ModuleCache>,
+        depth_limit: u32,
+    ) -> Self {
+        Self {
+            source,
+            cx,
+            cache,
+            in_progress: Mutex::new(Vec::new()),
+            depth: Mutex::new(0),
+            depth_limit,
+        }
+    }
+
+    /// Validate `ast` as a module body: only top-level `fn` decls,
+    /// `const` decls, and `import` statements are allowed. Top-level
+    /// expressions (which would execute on import — a footgun for
+    /// cache semantics) are rejected.
+    ///
+    /// `fn` declarations live in a separate slot on the AST and are
+    /// not in `statements()`, so the only allowed `Stmt` variants we
+    /// expect to see at top level are `Var` (when `CONSTANT` flag is
+    /// set) and `Import`. Anything else triggers a `ModuleShape` error.
+    fn check_module_shape(ast: &AST, name: &str) -> Result<(), String> {
+        use rhai::ASTFlags;
+        for stmt in ast.statements() {
+            match stmt {
+                rhai::Stmt::Var(_, opts, _) if opts.intersects(ASTFlags::CONSTANT) => {}
+                rhai::Stmt::Import(..) | rhai::Stmt::Noop(..) => {}
+                other => {
+                    return Err(format!(
+                        "module {name:?}: top-level {} is not allowed; \
+                         modules may only contain fn declarations, \
+                         const declarations, and import statements",
+                        stmt_kind_label(other),
+                    ));
+                }
+            }
+        }
+        Ok(())
+    }
+
+    /// Walk a compiled AST and collect the literal-path `import "<name>"`
+    /// declarations. Dynamic imports (e.g. `import some_var as y;`) are
+    /// skipped because the dep-graph can only track names known at
+    /// compile time. Exposed via [`extract_imports`] so the manager's
+    /// admin endpoints can populate the `script_imports` table from
+    /// the same logic the resolver uses.
+    fn extract_imports_inner(ast: &AST) -> Vec<String> {
+        let mut out = Vec::new();
+        for stmt in ast.statements() {
+            if let rhai::Stmt::Import(boxed, _) = stmt {
+                let (path_expr, _alias) = boxed.as_ref();
+                if let rhai::Expr::StringConstant(s, _) = path_expr {
+                    out.push(s.to_string());
+                }
+            }
+        }
+        out
+    }
+}
+
+/// Compile-and-validate a candidate module body. Public so the
+/// `Engine::validate_module` impl in `engine.rs` can call into it
+/// without duplicating the shape check.
+pub fn compile_module_ast(engine: &RhaiEngine, source: &str) -> Result<AST, String> {
+    let ast = engine.compile(source).map_err(|e| e.to_string())?;
+    PicloudModuleResolver::check_module_shape(&ast, "<source>")?;
+    Ok(ast)
+}
+
+/// Parse `source` as an endpoint script (no module-shape check) and
+/// return its declared literal-path imports. Used by
+/// `Engine::validate` to populate `ValidatedScript::imports` so the
+/// repo can write dep-graph edges.
+pub fn extract_imports(engine: &RhaiEngine, source: &str) -> Result<ValidatedScript, String> {
+    let ast = engine.compile(source).map_err(|e| e.to_string())?;
+    Ok(ValidatedScript {
+        imports: PicloudModuleResolver::extract_imports_inner(&ast),
+    })
+}
+
+/// Parse `source` as a module script: enforce shape, then extract
+/// imports. Used by `Engine::validate_module`.
+pub fn validate_module_source(
+    engine: &RhaiEngine,
+    source: &str,
+) -> Result<ValidatedScript, String> {
+    let ast = compile_module_ast(engine, source)?;
+    Ok(ValidatedScript {
+        imports: PicloudModuleResolver::extract_imports_inner(&ast),
+    })
+}
+
+fn stmt_kind_label(stmt: &rhai::Stmt) -> &'static str {
+    use rhai::ASTFlags;
+    match stmt {
+        rhai::Stmt::Var(_, opts, _) if opts.intersects(ASTFlags::CONSTANT) => "const declaration",
+        rhai::Stmt::Var(..) => "let declaration",
+        rhai::Stmt::Expr(..) => "expression",
+        rhai::Stmt::FnCall(..) => "function call",
+        rhai::Stmt::If(..) => "if statement",
+        rhai::Stmt::Switch(..) => "switch statement",
+        rhai::Stmt::While(..) => "while/loop statement",
+        rhai::Stmt::Do(..) => "do statement",
+        rhai::Stmt::For(..) => "for statement",
+        rhai::Stmt::Assignment(..) => "assignment",
+        rhai::Stmt::Block(..) => "block",
+        rhai::Stmt::TryCatch(..) => "try/catch",
+        rhai::Stmt::Return(..) => "return/throw statement",
+        rhai::Stmt::BreakLoop(..) => "break/continue",
+        rhai::Stmt::Import(..) => "import statement",
+        rhai::Stmt::Export(..) => "export statement",
+        _ => "statement",
+    }
+}
+
+impl ModuleResolver for PicloudModuleResolver {
+    #[allow(clippy::too_many_lines)]
+    fn resolve(
+        &self,
+        engine: &RhaiEngine,
+        _source: Option<&str>,
+        path: &str,
+        pos: Position,
+    ) -> Result<SharedRhaiModule, Box<EvalAltResult>> {
+        // RAII guard wraps both the depth counter and the import-stack
+        // push so that any early return (cycle / depth-exceeded / DB
+        // error / compile error / panic) leaves both consistent for
+        // any subsequent resolve() call on this resolver instance.
+        struct StackGuard<'r> {
+            stack: &'r Mutex<Vec<String>>,
+            depth: &'r Mutex<u32>,
+            armed: bool,
+        }
+        impl Drop for StackGuard<'_> {
+            fn drop(&mut self) {
+                if !self.armed {
+                    return;
+                }
+                if let Ok(mut s) = self.stack.lock() {
+                    s.pop();
+                }
+                if let Ok(mut d) = self.depth.lock() {
+                    *d = d.saturating_sub(1);
+                }
+            }
+        }
+
+        // Read-only check + atomic push under one lock pair, so a
+        // sibling resolve() call on a shared resolver instance can't
+        // race in between. (We don't expect parallel calls on the same
+        // resolver — Rhai evaluates a single AST on one thread — but
+        // grouping the operations is cheaper than reasoning about the
+        // future.)
+        {
+            let mut depth = self.depth.lock().expect("module depth lock poisoned");
+            if *depth >= self.depth_limit {
+                return Err(Box::new(EvalAltResult::ErrorInModule(
+                    path.to_string(),
+                    Box::new(EvalAltResult::ErrorRuntime(
+                        format!(
+                            "import depth limit ({}) exceeded while resolving {path:?}",
+                            self.depth_limit
+                        )
+                        .into(),
+                        pos,
+                    )),
+                    pos,
+                )));
+            }
+            let mut stack = self
+                .in_progress
+                .lock()
+                .expect("module in_progress lock poisoned");
+            if stack.iter().any(|p| p == path) {
+                let mut chain = stack.clone();
+                chain.push(path.to_string());
+                return Err(Box::new(EvalAltResult::ErrorInModule(
+                    path.to_string(),
+                    Box::new(EvalAltResult::ErrorRuntime(
+                        format!("circular import detected: {}", chain.join(" -> ")).into(),
+                        pos,
+                    )),
+                    pos,
+                )));
+            }
+            stack.push(path.to_string());
+            *depth += 1;
+        }
+        let _guard = StackGuard {
+            stack: &self.in_progress,
+            depth: &self.depth,
+            armed: true,
+        };
+
+        // Bridge to async. The resolver typically runs on a
+        // `spawn_blocking` thread (see LocalExecutorClient in
+        // orchestrator-core), but tests may invoke `Engine::execute`
+        // directly from a multi-threaded Tokio task. `try_current` +
+        // `block_in_place` covers both — on a blocking thread it's a
+        // no-op, on a worker thread it tells the runtime to relocate
+        // other tasks. `current_thread` runtimes still panic; non-
+        // Tokio contexts surface a clean Runtime error.
+        let handle = tokio::runtime::Handle::try_current().map_err(|_| {
+            Box::new(EvalAltResult::ErrorInModule(
+                path.to_string(),
+                Box::new(EvalAltResult::ErrorRuntime(
+                    "module resolver invoked outside a Tokio runtime; \
+                     wrap Engine::execute in tokio::task::spawn_blocking"
+                        .into(),
+                    pos,
+                )),
+                pos,
+            ))
+        })?;
+
+        let lookup_result: Result<Option<picloud_shared::ModuleScript>, ModuleSourceError> =
+            tokio::task::block_in_place(|| handle.block_on(self.source.lookup(&self.cx, path)));
+
+        let module_row = match lookup_result {
+            Ok(Some(m)) => m,
+            Ok(None) => {
+                return Err(Box::new(EvalAltResult::ErrorModuleNotFound(
+                    path.to_string(),
+                    pos,
+                )));
+            }
+            Err(e) => {
+                // v1.1.4 §10a: redact the backend error before it
+                // reaches a script. In public-HTTP context (principal:
+                // None) the verbatim message (e.g. "connection refused")
+                // leaks internal infrastructure shape. Log the original
+                // at error level for operators; surface a stable generic.
+                tracing::error!(
+                    target = "picloud::modules",
+                    app_id = %self.cx.app_id,
+                    module = path,
+                    error = %e,
+                    "module backend error"
+                );
+                return Err(Box::new(EvalAltResult::ErrorInModule(
+                    path.to_string(),
+                    Box::new(EvalAltResult::ErrorRuntime(
+                        "module backend unavailable; check server logs".into(),
+                        pos,
+                    )),
+                    pos,
+                )));
+            }
+        };
+
+        // Cache lookup: hit only if both key matches AND updated_at
+        // matches (cache is invalidated lazily on version change).
+        let cache_key = (self.cx.app_id, path.to_string());
+        {
+            let mut cache = self.cache.lock().expect("module cache lock poisoned");
+            if let Some(cached) = cache.get(&cache_key) {
+                if cached.updated_at == module_row.updated_at {
+                    tracing::debug!(
+                        target = "picloud::modules::cache",
+                        app_id = %self.cx.app_id,
+                        module = path,
+                        "cache hit"
+                    );
+                    return Ok(cached.module.clone());
+                }
+                tracing::debug!(
+                    target = "picloud::modules::cache",
+                    app_id = %self.cx.app_id,
+                    module = path,
+                    "cache stale; recompiling"
+                );
+            } else {
+                tracing::debug!(
+                    target = "picloud::modules::cache",
+                    app_id = %self.cx.app_id,
+                    module = path,
+                    "cache miss"
+                );
+            }
+        }
+
+        // Compile + module-shape validation. Module sources MAY have
+        // already been gated at create-time (admin endpoint runs
+        // `validate_module`), but we revalidate here to catch DB-direct
+        // inserts that bypass the API surface.
+        let ast = engine.compile(&module_row.source).map_err(|e| {
+            // Wrap as an ErrorRuntime to preserve the parse message
+            // text without trying to reconstruct rhai's internal
+            // ParseErrorType variant (which would require matching on
+            // its full variant set).
+            Box::new(EvalAltResult::ErrorInModule(
+                path.to_string(),
+                Box::new(EvalAltResult::ErrorRuntime(
+                    format!("module {path:?} parse error: {e}").into(),
+                    e.position(),
+                )),
+                pos,
+            ))
+        })?;
+
+        if let Err(msg) = Self::check_module_shape(&ast, path) {
+            return Err(Box::new(EvalAltResult::ErrorInModule(
+                path.to_string(),
+                Box::new(EvalAltResult::ErrorRuntime(msg.into(), pos)),
+                pos,
+            )));
+        }
+
+        // Rhai's eval_ast_as_new compiles the AST's body + functions
+        // into a Module that the importing script consumes via
+        // `path::fn(...)` calls. Recursive imports inside this module
+        // are resolved through the same `engine.set_module_resolver`
+        // (which is THIS resolver), so cycle/depth tracking carries
+        // through naturally.
+        let module = Module::eval_ast_as_new(rhai::Scope::new(), &ast, engine)
+            .map_err(|e| Box::new(EvalAltResult::ErrorInModule(path.to_string(), e, pos)))?;
+        let shared: SharedRhaiModule = module.into();
+
+        // Insert (possibly evicting via LRU). Subsequent imports of
+        // the same module under the same updated_at hit the cache.
+        {
+            let mut cache = self.cache.lock().expect("module cache lock poisoned");
+            cache.put(
+                cache_key,
+                CachedModule {
+                    updated_at: module_row.updated_at,
+                    module: shared.clone(),
+                },
+            );
+        }
+
+        Ok(shared)
+    }
+}

crates/executor-core/src/sandbox.rs

@@ -24,6 +24,12 @@ pub struct Limits {
     /// Max call/expression nesting depth.
     pub max_call_levels: usize,
     pub max_expr_depth: usize,
+
+    /// v1.1.3: hard ceiling on `import` chain depth (A→B→C→…). Independent
+    /// of cycle detection — guards against deep but acyclic graphs.
+    /// Not script-overridable (this is a platform-level guard, not a
+    /// per-script knob).
+    pub module_import_depth_max: u32,
 }
 
 impl Default for Limits {
@@ -35,6 +41,7 @@ impl Default for Limits {
             max_map_size: 10_000,
             max_call_levels: 64,
             max_expr_depth: 64,
+            module_import_depth_max: 8,
         }
     }
 }
@@ -65,6 +72,9 @@ impl Limits {
             max_expr_depth: overrides
                 .max_expr_depth
                 .map_or(self.max_expr_depth, narrow_usize),
+            // module_import_depth_max is platform-level — overrides
+            // never touch it. Carry through unchanged.
+            module_import_depth_max: self.module_import_depth_max,
         }
     }
 }

crates/executor-core/src/sdk/docs.rs

@@ -0,0 +1,255 @@
+//! `docs::` Rhai bridge — collection-scoped handle pattern, v1.1.2.
+//!
+//! ```rhai
+//! let users = docs::collection("users");
+//! let id = users.create(#{ name: "Alice", tier: "gold" });
+//! let doc = users.get(id);                 // envelope or () if missing
+//! let golds = users.find(#{ tier: "gold" });
+//! let one  = users.find_one(#{ tier: "gold" });
+//! users.update(id, #{ name: "Alice", tier: "platinum" });
+//! let removed = users.delete(id);           // bool was-present
+//! let page = users.list(#{ cursor: (), limit: 100 });
+//! ```
+//!
+//! Mirrors `kv.rs`: `DocsHandle` captures the collection + service +
+//! per-call cx; methods bind via `engine.register_fn` so scripts call
+//! them with dot-notation. **The service derives `app_id` from
+//! `cx.app_id` — never from any closure argument.** Cross-app
+//! isolation boundary; same as KV.
+//!
+//! Doc shape returned by `get`/`find`/`find_one`/`list`: an envelope
+//! `#{ id, data: #{...}, created_at, updated_at }`. Decision D in the
+//! v1.1.2 plan — explicit metadata vs user-data separation.
+
+use std::sync::Arc;
+
+use picloud_shared::{DocId, DocRow, DocsError, DocsService, SdkCallCx, Services};
+use rhai::{Array, Dynamic, Engine as RhaiEngine, EvalAltResult, Map, Module};
+use tokio::runtime::Handle as TokioHandle;
+use uuid::Uuid;
+
+use super::bridge::{dynamic_to_json, json_to_dynamic};
+
+/// Per-call handle captured by the Rhai SDK. Cheap to clone (two Arcs
+/// plus an owned string).
+#[derive(Clone)]
+pub struct DocsHandle {
+    collection: String,
+    service: Arc<dyn DocsService>,
+    cx: Arc<SdkCallCx>,
+}
+
+pub(super) fn register(engine: &mut RhaiEngine, services: &Services, cx: Arc<SdkCallCx>) {
+    let docs_service = services.docs.clone();
+
+    let mut module = Module::new();
+    {
+        let docs_service = docs_service.clone();
+        let cx = cx.clone();
+        module.set_native_fn(
+            "collection",
+            move |name: &str| -> Result<DocsHandle, Box<EvalAltResult>> {
+                if name.is_empty() {
+                    return Err("docs::collection name must not be empty".into());
+                }
+                Ok(DocsHandle {
+                    collection: name.to_string(),
+                    service: docs_service.clone(),
+                    cx: cx.clone(),
+                })
+            },
+        );
+    }
+    engine.register_static_module("docs", module.into());
+
+    engine.register_type_with_name::<DocsHandle>("DocsHandle");
+
+    register_create(engine);
+    register_get(engine);
+    register_find(engine);
+    register_find_one(engine);
+    register_update(engine);
+    register_delete(engine);
+    register_list(engine);
+}
+
+fn register_create(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "create",
+        |handle: &mut DocsHandle, data: Map| -> Result<String, Box<EvalAltResult>> {
+            let h = handle.clone();
+            let json = dynamic_to_json(&Dynamic::from(data));
+            let id = block_on(async move { h.service.create(&h.cx, &h.collection, json).await })?;
+            Ok(id.to_string())
+        },
+    );
+}
+
+fn register_get(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "get",
+        |handle: &mut DocsHandle, id: &str| -> Result<Dynamic, Box<EvalAltResult>> {
+            let h = handle.clone();
+            let parsed_id = parse_doc_id(id)?;
+            let row =
+                block_on(async move { h.service.get(&h.cx, &h.collection, parsed_id).await })?;
+            Ok(row.map_or(Dynamic::UNIT, |d| Dynamic::from(doc_to_map(&d))))
+        },
+    );
+}
+
+fn register_find(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "find",
+        |handle: &mut DocsHandle, filter: Map| -> Result<Array, Box<EvalAltResult>> {
+            let h = handle.clone();
+            let json = dynamic_to_json(&Dynamic::from(filter));
+            let rows = block_on(async move { h.service.find(&h.cx, &h.collection, json).await })?;
+            Ok(rows
+                .iter()
+                .map(|d| Dynamic::from(doc_to_map(d)))
+                .collect::<Vec<Dynamic>>())
+        },
+    );
+}
+
+fn register_find_one(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "find_one",
+        |handle: &mut DocsHandle, filter: Map| -> Result<Dynamic, Box<EvalAltResult>> {
+            let h = handle.clone();
+            let json = dynamic_to_json(&Dynamic::from(filter));
+            let row =
+                block_on(async move { h.service.find_one(&h.cx, &h.collection, json).await })?;
+            Ok(row.map_or(Dynamic::UNIT, |d| Dynamic::from(doc_to_map(&d))))
+        },
+    );
+}
+
+fn register_update(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "update",
+        |handle: &mut DocsHandle, id: &str, data: Map| -> Result<(), Box<EvalAltResult>> {
+            let h = handle.clone();
+            let parsed_id = parse_doc_id(id)?;
+            let json = dynamic_to_json(&Dynamic::from(data));
+            block_on(async move {
+                h.service
+                    .update(&h.cx, &h.collection, parsed_id, json)
+                    .await
+            })
+        },
+    );
+}
+
+fn register_delete(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "delete",
+        |handle: &mut DocsHandle, id: &str| -> Result<bool, Box<EvalAltResult>> {
+            let h = handle.clone();
+            let parsed_id = parse_doc_id(id)?;
+            block_on(async move { h.service.delete(&h.cx, &h.collection, parsed_id).await })
+        },
+    );
+}
+
+fn register_list(engine: &mut RhaiEngine) {
+    // Zero-arg form: full page from the start.
+    engine.register_fn(
+        "list",
+        |handle: &mut DocsHandle| -> Result<Map, Box<EvalAltResult>> { list_call(handle, None, 0) },
+    );
+    // One-arg form: pass `#{ cursor, limit }` map. Either field is
+    // optional; missing/unit → defaults.
+    engine.register_fn(
+        "list",
+        |handle: &mut DocsHandle, args: Map| -> Result<Map, Box<EvalAltResult>> {
+            let cursor = match args.get("cursor") {
+                Some(d) if !d.is_unit() => {
+                    Some(d.clone().into_string().map_err(|_| -> Box<EvalAltResult> {
+                        "docs::list: 'cursor' must be a string or ()".into()
+                    })?)
+                }
+                _ => None,
+            };
+            let limit = match args.get("limit") {
+                Some(d) if !d.is_unit() => {
+                    let n = d.as_int().map_err(|_| -> Box<EvalAltResult> {
+                        "docs::list: 'limit' must be an integer".into()
+                    })?;
+                    u32::try_from(n.max(0)).unwrap_or(0)
+                }
+                _ => 0,
+            };
+            list_call(handle, cursor, limit)
+        },
+    );
+}
+
+fn list_call(
+    handle: &DocsHandle,
+    cursor: Option<String>,
+    limit: u32,
+) -> Result<Map, Box<EvalAltResult>> {
+    let h = handle.clone();
+    let page = block_on(async move {
+        h.service
+            .list(&h.cx, &h.collection, cursor.as_deref(), limit)
+            .await
+    })?;
+    let mut m = Map::new();
+    let docs: Array = page
+        .docs
+        .iter()
+        .map(|d| Dynamic::from(doc_to_map(d)))
+        .collect();
+    m.insert("docs".into(), docs.into());
+    m.insert(
+        "next_cursor".into(),
+        page.next_cursor.map_or(Dynamic::UNIT, Dynamic::from),
+    );
+    Ok(m)
+}
+
+/// Build the `{ id, data, created_at, updated_at }` envelope per
+/// Decision D. Scripts read user fields via `doc.data.<field>`; `id`
+/// and timestamps are direct children of the envelope.
+fn doc_to_map(doc: &DocRow) -> Map {
+    let mut m = Map::new();
+    m.insert("id".into(), doc.id.to_string().into());
+    m.insert("data".into(), json_to_dynamic(doc.data.clone()));
+    m.insert("created_at".into(), doc.created_at.to_rfc3339().into());
+    m.insert("updated_at".into(), doc.updated_at.to_rfc3339().into());
+    m
+}
+
+fn parse_doc_id(id: &str) -> Result<DocId, Box<EvalAltResult>> {
+    Uuid::parse_str(id).map_err(|e| -> Box<EvalAltResult> {
+        EvalAltResult::ErrorRuntime(
+            format!("docs: invalid id '{id}': {e}").into(),
+            rhai::Position::NONE,
+        )
+        .into()
+    })
+}
+
+/// Mirrors `kv.rs::block_on` — Tokio runtime is reachable from inside
+/// the `spawn_blocking` wrapper that owns Rhai execution. Errors
+/// prefix with `"docs: "` so scripts see `docs: forbidden`,
+/// `docs: document not found`, `docs: unsupported operator: …`, etc.
+fn block_on<F, T>(fut: F) -> Result<T, Box<EvalAltResult>>
+where
+    F: std::future::Future<Output = Result<T, DocsError>> + Send,
+    T: Send,
+{
+    let handle = TokioHandle::try_current().map_err(|e| -> Box<EvalAltResult> {
+        EvalAltResult::ErrorRuntime(
+            format!("docs: no tokio runtime available: {e}").into(),
+            rhai::Position::NONE,
+        )
+        .into()
+    })?;
+    handle.block_on(fut).map_err(|err| -> Box<EvalAltResult> {
+        EvalAltResult::ErrorRuntime(format!("docs: {err}").into(), rhai::Position::NONE).into()
+    })
+}

crates/executor-core/src/sdk/files.rs

@@ -0,0 +1,281 @@
+//! `files::` Rhai bridge — collection-scoped handle pattern (v1.1.5).
+//!
+//! ```rhai
+//! let avatars = files::collection("avatars");
+//! let id = avatars.create(#{ name: "a.jpg", content_type: "image/jpeg", data: blob });
+//! let meta  = avatars.head(id);   // metadata map or ()
+//! let bytes = avatars.get(id);    // Blob or ()
+//! avatars.update(id, #{ data: new_bytes });
+//! let gone  = avatars.delete(id); // bool (was-present)
+//! let page  = avatars.list();     // #{ files: [...], next_cursor: () }
+//! ```
+//!
+//! The `FilesHandle` custom Rhai type captures the collection name once
+//! and routes each call through the injected `Arc<dyn FilesService>`
+//! with the per-call `Arc<SdkCallCx>`. **The service derives `app_id`
+//! from `cx.app_id` — it never appears in any signature script-side,
+//! preserving cross-app isolation.**
+//!
+//! Error convention (per `docs/sdk-shape.md`): `create`/`update`/
+//! `delete` throw on failure; `get`/`head` return `()` for a missing
+//! file; `delete` returns `bool` (was-present). The blob bytes are a
+//! Rhai `Blob` (byte array) in both directions.
+
+use std::sync::Arc;
+
+use picloud_shared::{
+    FileMeta, FileUpdate, FilesError, FilesService, NewFile, SdkCallCx, Services,
+};
+use rhai::{Array, Dynamic, Engine as RhaiEngine, EvalAltResult, Map, Module};
+use tokio::runtime::Handle as TokioHandle;
+
+/// Per-call handle captured by the Rhai SDK. Cheap to clone (two Arcs
+/// plus an owned string).
+#[derive(Clone)]
+pub struct FilesHandle {
+    collection: String,
+    service: Arc<dyn FilesService>,
+    cx: Arc<SdkCallCx>,
+}
+
+pub(super) fn register(engine: &mut RhaiEngine, services: &Services, cx: Arc<SdkCallCx>) {
+    let files_service = services.files.clone();
+
+    let mut module = Module::new();
+    {
+        let files_service = files_service.clone();
+        let cx = cx.clone();
+        module.set_native_fn(
+            "collection",
+            move |name: &str| -> Result<FilesHandle, Box<EvalAltResult>> {
+                if name.is_empty() {
+                    return Err("files::collection name must not be empty".into());
+                }
+                Ok(FilesHandle {
+                    collection: name.to_string(),
+                    service: files_service.clone(),
+                    cx: cx.clone(),
+                })
+            },
+        );
+    }
+    engine.register_static_module("files", module.into());
+
+    engine.register_type_with_name::<FilesHandle>("FilesHandle");
+
+    register_create(engine);
+    register_head(engine);
+    register_get(engine);
+    register_update(engine);
+    register_delete(engine);
+    register_list(engine);
+}
+
+fn register_create(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "create",
+        |handle: &mut FilesHandle, meta: Map| -> Result<String, Box<EvalAltResult>> {
+            let name = require_string(&meta, "name")?;
+            let content_type = require_string(&meta, "content_type")?;
+            let data = require_blob(&meta, "data")?;
+            let h = handle.clone();
+            let new = NewFile {
+                name,
+                content_type,
+                data,
+            };
+            let id = block_on(async move { h.service.create(&h.cx, &h.collection, new).await })?;
+            Ok(id.to_string())
+        },
+    );
+}
+
+fn register_head(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "head",
+        |handle: &mut FilesHandle, id: &str| -> Result<Dynamic, Box<EvalAltResult>> {
+            let h = handle.clone();
+            let id = id.to_string();
+            let meta = block_on(async move { h.service.head(&h.cx, &h.collection, &id).await })?;
+            Ok(meta.map_or(Dynamic::UNIT, |m| file_meta_to_map(&m).into()))
+        },
+    );
+}
+
+fn register_get(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "get",
+        |handle: &mut FilesHandle, id: &str| -> Result<Dynamic, Box<EvalAltResult>> {
+            let h = handle.clone();
+            let id = id.to_string();
+            let bytes = block_on(async move { h.service.get(&h.cx, &h.collection, &id).await })?;
+            Ok(bytes.map_or(Dynamic::UNIT, Dynamic::from_blob))
+        },
+    );
+}
+
+fn register_update(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "update",
+        |handle: &mut FilesHandle, id: &str, meta: Map| -> Result<(), Box<EvalAltResult>> {
+            let data = require_blob(&meta, "data")?;
+            let name = optional_string(&meta, "name")?;
+            let content_type = optional_string(&meta, "content_type")?;
+            let h = handle.clone();
+            let id = id.to_string();
+            let upd = FileUpdate {
+                data,
+                name,
+                content_type,
+            };
+            block_on(async move { h.service.update(&h.cx, &h.collection, &id, upd).await })
+        },
+    );
+}
+
+fn register_delete(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "delete",
+        |handle: &mut FilesHandle, id: &str| -> Result<bool, Box<EvalAltResult>> {
+            let h = handle.clone();
+            let id = id.to_string();
+            block_on(async move { h.service.delete(&h.cx, &h.collection, &id).await })
+        },
+    );
+}
+
+fn register_list(engine: &mut RhaiEngine) {
+    engine.register_fn(
+        "list",
+        |handle: &mut FilesHandle| -> Result<Map, Box<EvalAltResult>> {
+            list_call(handle, None, 0)
+        },
+    );
+    engine.register_fn(
+        "list",
+        |handle: &mut FilesHandle, cursor: &str| -> Result<Map, Box<EvalAltResult>> {
+            list_call(handle, Some(cursor.to_string()), 0)
+        },
+    );
+    engine.register_fn(
+        "list",
+        |handle: &mut FilesHandle, cursor: &str, limit: i64| -> Result<Map, Box<EvalAltResult>> {
+            let limit = u32::try_from(limit.max(0)).unwrap_or(0);
+            list_call(handle, Some(cursor.to_string()), limit)
+        },
+    );
+    // `list(#{ cursor, limit })` — the map form documented in the brief.
+    engine.register_fn(
+        "list",
+        |handle: &mut FilesHandle, opts: Map| -> Result<Map, Box<EvalAltResult>> {
+            let cursor = match opts.get("cursor") {
+                Some(v) if !v.is_unit() => {
+                    Some(v.clone().into_string().map_err(|_| -> Box<EvalAltResult> {
+                        "files: list cursor must be a string".into()
+                    })?)
+                }
+                _ => None,
+            };
+            let limit = match opts.get("limit") {
+                Some(v) if !v.is_unit() => {
+                    u32::try_from(v.as_int().unwrap_or(0).max(0)).unwrap_or(0)
+                }
+                _ => 0,
+            };
+            list_call(handle, cursor, limit)
+        },
+    );
+}
+
+fn list_call(
+    handle: &FilesHandle,
+    cursor: Option<String>,
+    limit: u32,
+) -> Result<Map, Box<EvalAltResult>> {
+    let h = handle.clone();
+    let page = block_on(async move {
+        h.service
+            .list(&h.cx, &h.collection, cursor.as_deref(), limit)
+            .await
+    })?;
+    let mut m = Map::new();
+    let files: Array = page
+        .files
+        .iter()
+        .map(|meta| Dynamic::from(file_meta_to_map(meta)))
+        .collect();
+    m.insert("files".into(), files.into());
+    m.insert(
+        "next_cursor".into(),
+        page.next_cursor.map_or(Dynamic::UNIT, Dynamic::from),
+    );
+    Ok(m)
+}
+
+/// Render a `FileMeta` into the Rhai map shape scripts see from
+/// `head` / `list`.
+fn file_meta_to_map(meta: &FileMeta) -> Map {
+    let mut m = Map::new();
+    m.insert("id".into(), meta.id.to_string().into());
+    m.insert("collection".into(), meta.collection.clone().into());
+    m.insert("name".into(), meta.name.clone().into());
+    m.insert("content_type".into(), meta.content_type.clone().into());
+    m.insert(
+        "size".into(),
+        i64::try_from(meta.size).unwrap_or(i64::MAX).into(),
+    );
+    m.insert("checksum".into(), meta.checksum.clone().into());
+    m.insert("created_at".into(), meta.created_at.to_rfc3339().into());
+    m.insert("updated_at".into(), meta.updated_at.to_rfc3339().into());
+    m
+}
+
+/// Pull a required string field out of a Rhai map; throw naming the
+/// field if it's absent or not a string.
+fn require_string(meta: &Map, field: &'static str) -> Result<String, Box<EvalAltResult>> {
+    match meta.get(field) {
+        Some(v) if v.is_string() => Ok(v.clone().into_string().unwrap_or_default()),
+        Some(_) => Err(format!("files::create: field '{field}' must be a string").into()),
+        None => Err(format!("files::create: missing required field '{field}'").into()),
+    }
+}
+
+/// Pull an optional string field; `None` when the key is absent or unit.
+fn optional_string(meta: &Map, field: &'static str) -> Result<Option<String>, Box<EvalAltResult>> {
+    match meta.get(field) {
+        None => Ok(None),
+        Some(v) if v.is_unit() => Ok(None),
+        Some(v) if v.is_string() => Ok(Some(v.clone().into_string().unwrap_or_default())),
+        Some(_) => Err(format!("files::update: field '{field}' must be a string").into()),
+    }
+}
+
+/// Pull a required blob (`data`) out of a Rhai map; throw naming the
+/// field if it's absent or not a blob.
+fn require_blob(meta: &Map, field: &'static str) -> Result<Vec<u8>, Box<EvalAltResult>> {
+    match meta.get(field) {
+        Some(v) if v.is_blob() => Ok(v.clone().into_blob().unwrap_or_default()),
+        Some(_) => Err(format!("files: field '{field}' must be a Blob (byte array)").into()),
+        None => Err(format!("files: missing required field '{field}'").into()),
+    }
+}
+
+/// Run an async future inside the synchronous Rhai context. Mirrors
+/// `kv::block_on`; safe because `LocalExecutorClient` runs the script
+/// under `spawn_blocking`, so a runtime handle is reachable.
+fn block_on<F, T>(fut: F) -> Result<T, Box<EvalAltResult>>
+where
+    F: std::future::Future<Output = Result<T, FilesError>> + Send,
+    T: Send,
+{
+    let handle = TokioHandle::try_current().map_err(|e| -> Box<EvalAltResult> {
+        EvalAltResult::ErrorRuntime(
+            format!("files: no tokio runtime available: {e}").into(),
+            rhai::Position::NONE,
+        )
+        .into()
+    })?;
+    handle.block_on(fut).map_err(|err| -> Box<EvalAltResult> {
+        EvalAltResult::ErrorRuntime(format!("files: {err}").into(), rhai::Position::NONE).into()
+    })
+}

crates/executor-core/src/sdk/http.rs

@@ -0,0 +1,391 @@
+//! `http::` Rhai bridge — outbound HTTP from scripts (v1.1.4).
+//!
+//! ```rhai
+//! let r = http::get("https://api.example.com/users/123");
+//! let r = http::get(url, #{ headers: #{ "Authorization": "Bearer x" }, timeout_ms: 5000 });
+//! let r = http::post(url, #{ text: "hello" });          // Map body → JSON
+//! let r = http::post(url, "raw", #{ headers: #{ ... } }); // String body → text/plain
+//! let r = http::post_form(url, #{ a: "1", b: "2" });    // form-encoded
+//! let r = http::request("OPTIONS", url);
+//! ```
+//!
+//! **Argument shape (v1.1.4 decision):** body and options are separate
+//! positional arguments — `verb(url, body, opts)` — not body-inside-
+//! opts. This keeps the unknown-opt-key typo guard intact and resolves
+//! the brief's internal contradiction (its Slack example passed a bare
+//! body map). The `opts` vocabulary is exactly
+//! `{headers, timeout_ms, follow_redirects, max_redirects}`; any other
+//! key throws.
+//!
+//! Body dispatch (positional `body`): Map/Array → JSON +
+//! `application/json`; String → raw + `text/plain`; Unit `()` → no
+//! body. GET/HEAD ignore any body.
+//!
+//! Response is a Rhai map `#{ status, headers, body, body_raw }`:
+//! `body` is the parsed JSON when the response is `application/json`
+//! and parses; `()` for an empty body; otherwise the raw string.
+//!
+//! Errors follow `docs/sdk-shape.md`: network/timeout/SSRF/size failures
+//! throw (`"http: <message>"`); a non-2xx status does NOT throw — the
+//! response map is returned, fetch-style.
+
+use std::collections::BTreeMap;
+use std::sync::Arc;
+
+use picloud_shared::{HttpError, HttpRequest, HttpResponse, HttpService, SdkCallCx, Services};
+use rhai::{Dynamic, Engine as RhaiEngine, EvalAltResult, Map, Module};
+use tokio::runtime::Handle as TokioHandle;
+
+use super::bridge::{dynamic_to_json, json_to_dynamic};
+
+/// Bridge-side defaults (the service clamps server-side too). The
+/// `MAX_*` ceilings stay `i64` because they're compared against the
+/// raw `i64` the script passed (so an over-limit value is rejected, not
+/// truncated); the defaults are `u32` to match the `Opts` fields.
+const DEFAULT_TIMEOUT_MS: u32 = 30_000;
+const MAX_TIMEOUT_MS: i64 = 60_000;
+const DEFAULT_MAX_REDIRECTS: u32 = 5;
+const MAX_REDIRECTS: i64 = 10;
+
+const ALLOWED_OPT_KEYS: [&str; 4] = ["headers", "timeout_ms", "follow_redirects", "max_redirects"];
+
+pub(super) fn register(engine: &mut RhaiEngine, services: &Services, cx: Arc<SdkCallCx>) {
+    let svc = services.http.clone();
+    let mut module = Module::new();
+
+    // Bodyless verbs: (url) / (url, opts).
+    for verb in ["get", "head"] {
+        register_bodyless(&mut module, verb, &svc, &cx);
+    }
+    // Body verbs: (url) / (url, body) / (url, body, opts).
+    for verb in ["post", "put", "patch", "delete"] {
+        register_body(&mut module, verb, &svc, &cx);
+    }
+    register_post_form(&mut module, &svc, &cx);
+    register_request(&mut module, &svc, &cx);
+
+    engine.register_static_module("http", module.into());
+}
+
+fn register_bodyless(
+    module: &mut Module,
+    verb: &'static str,
+    svc: &Arc<dyn HttpService>,
+    cx: &Arc<SdkCallCx>,
+) {
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn(verb, move |url: &str| {
+            invoke(&svc, &cx, verb, url, None, None)
+        });
+    }
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn(verb, move |url: &str, opts: Map| {
+            invoke(&svc, &cx, verb, url, None, Some(&opts))
+        });
+    }
+}
+
+fn register_body(
+    module: &mut Module,
+    verb: &'static str,
+    svc: &Arc<dyn HttpService>,
+    cx: &Arc<SdkCallCx>,
+) {
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn(verb, move |url: &str| {
+            invoke(&svc, &cx, verb, url, None, None)
+        });
+    }
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn(verb, move |url: &str, body: Dynamic| {
+            invoke(&svc, &cx, verb, url, Some(body), None)
+        });
+    }
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn(verb, move |url: &str, body: Dynamic, opts: Map| {
+            invoke(&svc, &cx, verb, url, Some(body), Some(&opts))
+        });
+    }
+}
+
+fn register_post_form(module: &mut Module, svc: &Arc<dyn HttpService>, cx: &Arc<SdkCallCx>) {
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn("post_form", move |url: &str, form: Map| {
+            invoke_form(&svc, &cx, url, &form, None)
+        });
+    }
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn("post_form", move |url: &str, form: Map, opts: Map| {
+            invoke_form(&svc, &cx, url, &form, Some(&opts))
+        });
+    }
+}
+
+fn register_request(module: &mut Module, svc: &Arc<dyn HttpService>, cx: &Arc<SdkCallCx>) {
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn("request", move |method: &str, url: &str| {
+            invoke(&svc, &cx, method, url, None, None)
+        });
+    }
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn("request", move |method: &str, url: &str, body: Dynamic| {
+            invoke(&svc, &cx, method, url, Some(body), None)
+        });
+    }
+    {
+        let (svc, cx) = (svc.clone(), cx.clone());
+        module.set_native_fn(
+            "request",
+            move |method: &str, url: &str, body: Dynamic, opts: Map| {
+                invoke(&svc, &cx, method, url, Some(body), Some(&opts))
+            },
+        );
+    }
+}
+
+/// Parsed `opts` map.
+struct Opts {
+    headers: BTreeMap<String, String>,
+    timeout_ms: u32,
+    follow_redirects: bool,
+    max_redirects: u32,
+}
+
+impl Default for Opts {
+    fn default() -> Self {
+        Self {
+            headers: BTreeMap::new(),
+            timeout_ms: DEFAULT_TIMEOUT_MS,
+            follow_redirects: true,
+            max_redirects: DEFAULT_MAX_REDIRECTS,
+        }
+    }
+}
+
+fn parse_opts(opts: Option<&Map>) -> Result<Opts, Box<EvalAltResult>> {
+    let mut out = Opts::default();
+    let Some(map) = opts else {
+        return Ok(out);
+    };
+    for key in map.keys() {
+        if !ALLOWED_OPT_KEYS.contains(&key.as_str()) {
+            return Err(err(format!("unknown option key: {key}")));
+        }
+    }
+    if let Some(h) = map.get("headers") {
+        let hm = h
+            .clone()
+            .try_cast::<Map>()
+            .ok_or_else(|| err("headers must be a map".to_string()))?;
+        for (k, v) in hm {
+            out.headers.insert(k.to_string(), dyn_to_string(&v));
+        }
+    }
+    if let Some(t) = map.get("timeout_ms") {
+        let ms = t
+            .as_int()
+            .map_err(|_| err("timeout_ms must be an integer".to_string()))?;
+        if ms > MAX_TIMEOUT_MS {
+            return Err(err(format!(
+                "timeout_ms {ms} exceeds the {MAX_TIMEOUT_MS}ms maximum"
+            )));
+        }
+        if ms > 0 {
+            out.timeout_ms = u32::try_from(ms).unwrap_or(u32::MAX);
+        }
+    }
+    if let Some(f) = map.get("follow_redirects") {
+        out.follow_redirects = f
+            .as_bool()
+            .map_err(|_| err("follow_redirects must be a bool".to_string()))?;
+    }
+    if let Some(m) = map.get("max_redirects") {
+        let n = m
+            .as_int()
+            .map_err(|_| err("max_redirects must be an integer".to_string()))?;
+        if n > MAX_REDIRECTS {
+            return Err(err(format!(
+                "max_redirects {n} exceeds the {MAX_REDIRECTS} maximum"
+            )));
+        }
+        out.max_redirects = u32::try_from(n.max(0)).unwrap_or(0);
+    }
+    Ok(out)
+}
+
+/// Encoded request body + the content-type chosen for it.
+type EncodedBody = (Option<Vec<u8>>, Option<String>);
+
+/// Dispatch a positional body by Rhai type. Returns the encoded bytes +
+/// the chosen content-type. GET/HEAD callers pass `body = None`, so
+/// this is never reached for them.
+fn dispatch_body(body: Dynamic) -> Result<EncodedBody, Box<EvalAltResult>> {
+    if body.is_unit() {
+        return Ok((None, None));
+    }
+    if body.is_string() {
+        let s = body.into_string().unwrap_or_default();
+        return Ok((Some(s.into_bytes()), Some("text/plain".to_string())));
+    }
+    if body.is_map() || body.is_array() {
+        let json = dynamic_to_json(&body);
+        let bytes = serde_json::to_vec(&json)
+            .map_err(|e| err(format!("could not encode JSON body: {e}")))?;
+        return Ok((Some(bytes), Some("application/json".to_string())));
+    }
+    // Scalars (int/float/bool) → JSON-encode for consistency.
+    let json = dynamic_to_json(&body);
+    let bytes =
+        serde_json::to_vec(&json).map_err(|e| err(format!("could not encode body: {e}")))?;
+    Ok((Some(bytes), Some("application/json".to_string())))
+}
+
+#[allow(clippy::needless_pass_by_value)]
+fn invoke(
+    svc: &Arc<dyn HttpService>,
+    cx: &Arc<SdkCallCx>,
+    method: &str,
+    url: &str,
+    body: Option<Dynamic>,
+    opts: Option<&Map>,
+) -> Result<Dynamic, Box<EvalAltResult>> {
+    let opts = parse_opts(opts)?;
+    let method_uc = method.to_ascii_uppercase();
+    let bodyless = matches!(method_uc.as_str(), "GET" | "HEAD");
+    let (encoded, content_type) = if bodyless {
+        (None, None)
+    } else if let Some(b) = body {
+        dispatch_body(b)?
+    } else {
+        (None, None)
+    };
+
+    let req = HttpRequest {
+        method: method_uc,
+        url: url.to_string(),
+        headers: opts.headers,
+        body: encoded,
+        content_type,
+        timeout_ms: opts.timeout_ms,
+        follow_redirects: opts.follow_redirects,
+        max_redirects: opts.max_redirects,
+        script_id: Some(cx.script_id.to_string()),
+    };
+    let resp = block_on(svc, cx, req)?;
+    Ok(response_to_dynamic(&resp))
+}
+
+#[allow(clippy::needless_pass_by_value)]
+fn invoke_form(
+    svc: &Arc<dyn HttpService>,
+    cx: &Arc<SdkCallCx>,
+    url: &str,
+    form: &Map,
+    opts: Option<&Map>,
+) -> Result<Dynamic, Box<EvalAltResult>> {
+    let opts = parse_opts(opts)?;
+    let mut serializer = url::form_urlencoded::Serializer::new(String::new());
+    for (k, v) in form {
+        serializer.append_pair(k.as_str(), &dyn_to_string(v));
+    }
+    let encoded = serializer.finish();
+
+    let req = HttpRequest {
+        method: "POST".to_string(),
+        url: url.to_string(),
+        headers: opts.headers,
+        body: Some(encoded.into_bytes()),
+        content_type: Some("application/x-www-form-urlencoded".to_string()),
+        timeout_ms: opts.timeout_ms,
+        follow_redirects: opts.follow_redirects,
+        max_redirects: opts.max_redirects,
+        script_id: Some(cx.script_id.to_string()),
+    };
+    let resp = block_on(svc, cx, req)?;
+    Ok(response_to_dynamic(&resp))
+}
+
+fn response_to_dynamic(resp: &HttpResponse) -> Dynamic {
+    let mut m = Map::new();
+    m.insert("status".into(), i64::from(resp.status).into());
+
+    let mut headers = Map::new();
+    let mut content_type = String::new();
+    for (k, v) in &resp.headers {
+        if k == "content-type" {
+            content_type.clone_from(v);
+        }
+        headers.insert(k.clone().into(), v.clone().into());
+    }
+    m.insert("headers".into(), headers.into());
+
+    // `body`: parsed JSON when the response is JSON and parses; () when
+    // empty; otherwise the raw string.
+    let body = if resp.body_raw.is_empty() {
+        Dynamic::UNIT
+    } else if content_type
+        .to_ascii_lowercase()
+        .starts_with("application/json")
+    {
+        match serde_json::from_str::<serde_json::Value>(&resp.body_raw) {
+            Ok(json) => json_to_dynamic(json),
+            Err(_) => resp.body_raw.clone().into(),
+        }
+    } else {
+        resp.body_raw.clone().into()
+    };
+    m.insert("body".into(), body);
+    m.insert("body_raw".into(), resp.body_raw.clone().into());
+    m.into()
+}
+
+fn dyn_to_string(v: &Dynamic) -> String {
+    if v.is_string() {
+        v.clone().into_string().unwrap_or_default()
+    } else {
+        v.to_string()
+    }
+}
+
+// Rhai's native-fn error channel is `Box<EvalAltResult>`, so these
+// helpers return the boxed form the call sites need.
+#[allow(clippy::unnecessary_box_returns)]
+fn err(msg: String) -> Box<EvalAltResult> {
+    EvalAltResult::ErrorRuntime(format!("http: {msg}").into(), rhai::Position::NONE).into()
+}
+
+/// Run the async service call from the synchronous Rhai context. Same
+/// pattern as `kv`/`docs`: the script runs under `spawn_blocking`, so a
+/// runtime handle is reachable and blocking on it is correct.
+fn block_on(
+    svc: &Arc<dyn HttpService>,
+    cx: &Arc<SdkCallCx>,
+    req: HttpRequest,
+) -> Result<HttpResponse, Box<EvalAltResult>> {
+    let handle = TokioHandle::try_current().map_err(|e| -> Box<EvalAltResult> {
+        EvalAltResult::ErrorRuntime(
+            format!("http: no tokio runtime available: {e}").into(),
+            rhai::Position::NONE,
+        )
+        .into()
+    })?;
+    let svc = svc.clone();
+    let cx = cx.clone();
+    handle
+        .block_on(async move { svc.request(&cx, req).await })
+        .map_err(map_http_err)
+}
+
+#[allow(clippy::unnecessary_box_returns)]
+fn map_http_err(e: HttpError) -> Box<EvalAltResult> {
+    EvalAltResult::ErrorRuntime(format!("http: {e}").into(), rhai::Position::NONE).into()
+}

crates/executor-core/src/sdk/mod.rs

@@ -14,7 +14,11 @@
 pub mod bridge;
 pub mod cx;
 pub mod dead_letters;
+pub mod docs;
+pub mod files;
+pub mod http;
 pub mod kv;
+pub mod pubsub;
 pub mod stdlib;
 
 pub use bridge::{dynamic_to_json, json_to_dynamic};
@@ -33,5 +37,9 @@ use rhai::Engine as RhaiEngine;
 /// single `<service>::register(...)` line per service.
 pub fn register_all(engine: &mut RhaiEngine, services: &Services, cx: Arc<SdkCallCx>) {
     kv::register(engine, services, cx.clone());
-    dead_letters::register(engine, services, cx);
+    docs::register(engine, services, cx.clone());
+    dead_letters::register(engine, services, cx.clone());
+    http::register(engine, services, cx.clone());
+    files::register(engine, services, cx.clone());
+    pubsub::register(engine, services, cx);
 }

crates/executor-core/src/sdk/pubsub.rs

@@ -0,0 +1,100 @@
+//! `pubsub::` Rhai bridge — durable publish (v1.1.5).
+//!
+//! ```rhai
+//! pubsub::publish_durable("user.created", #{ user_id: "abc" });
+//! pubsub::publish_durable("metric", 42);
+//! ```
+//!
+//! No handle pattern (topics ARE the grouping unit, so there's no
+//! `::collection(...)`). The message is any JSON-serializable Rhai value
+//! — Maps, Arrays, strings, numbers, bools, unit, and **Blobs (which
+//! encode as base64 strings** so trigger handlers see them as base64 on
+//! the wire). Nested blobs are encoded at any depth.
+//!
+//! `app_id` is derived from `cx.app_id` in the service — it never
+//! appears in the script-side signature, preserving cross-app
+//! isolation.
+
+use std::sync::Arc;
+
+use base64::engine::general_purpose::STANDARD;
+use base64::Engine as _;
+use picloud_shared::{PubsubError, SdkCallCx, Services};
+use rhai::{Array, Dynamic, Engine as RhaiEngine, EvalAltResult, Map, Module};
+use serde_json::Value as Json;
+use tokio::runtime::Handle as TokioHandle;
+
+pub(super) fn register(engine: &mut RhaiEngine, services: &Services, cx: Arc<SdkCallCx>) {
+    let svc = services.pubsub.clone();
+    let mut module = Module::new();
+    {
+        let svc = svc.clone();
+        let cx = cx.clone();
+        module.set_native_fn(
+            "publish_durable",
+            move |topic: &str, message: Dynamic| -> Result<(), Box<EvalAltResult>> {
+                let json = message_to_json(&message);
+                let svc = svc.clone();
+                let cx = cx.clone();
+                block_on(async move { svc.publish_durable(&cx, topic, json).await })
+            },
+        );
+    }
+    engine.register_static_module("pubsub", module.into());
+}
+
+/// Convert a Rhai `Dynamic` message into JSON, base64-encoding any
+/// `Blob` (at any nesting depth). Mirrors `bridge::dynamic_to_json` but
+/// adds the blob arm the pub/sub wire contract requires.
+fn message_to_json(value: &Dynamic) -> Json {
+    // Blob must be checked before the generic array path (a Blob is a
+    // `Vec<u8>`, distinct from a Rhai `Array`).
+    if value.is_blob() {
+        let blob = value.clone().into_blob().unwrap_or_default();
+        return Json::String(STANDARD.encode(&blob));
+    }
+    if value.is_unit() {
+        return Json::Null;
+    }
+    if let Ok(b) = value.as_bool() {
+        return Json::Bool(b);
+    }
+    if let Ok(i) = value.as_int() {
+        return Json::Number(i.into());
+    }
+    if let Ok(f) = value.as_float() {
+        return serde_json::Number::from_f64(f).map_or(Json::Null, Json::Number);
+    }
+    if value.is_string() {
+        return Json::String(value.clone().into_string().unwrap_or_default());
+    }
+    if let Some(arr) = value.clone().try_cast::<Array>() {
+        return Json::Array(arr.iter().map(message_to_json).collect());
+    }
+    if let Some(map) = value.clone().try_cast::<Map>() {
+        let mut out = serde_json::Map::new();
+        for (k, v) in map {
+            out.insert(k.to_string(), message_to_json(&v));
+        }
+        return Json::Object(out);
+    }
+    Json::String(value.to_string())
+}
+
+/// Run an async future inside the synchronous Rhai context. Mirrors
+/// `kv::block_on`.
+fn block_on<F>(fut: F) -> Result<(), Box<EvalAltResult>>
+where
+    F: std::future::Future<Output = Result<(), PubsubError>> + Send,
+{
+    let handle = TokioHandle::try_current().map_err(|e| -> Box<EvalAltResult> {
+        EvalAltResult::ErrorRuntime(
+            format!("pubsub: no tokio runtime available: {e}").into(),
+            rhai::Position::NONE,
+        )
+        .into()
+    })?;
+    handle.block_on(fut).map_err(|err| -> Box<EvalAltResult> {
+        EvalAltResult::ErrorRuntime(format!("pubsub: {err}").into(), rhai::Position::NONE).into()
+    })
+}

crates/executor-core/tests/module_redaction_logging.rs

@@ -0,0 +1,129 @@
+//! v1.1.4 §10a: the original module backend error MUST be logged at
+//! error level (so operators can still diagnose), even though it is
+//! redacted from the script-visible error.
+//!
+//! This test owns the process-global tracing subscriber, so it lives in
+//! its own integration-test binary (one `set_global_default` per
+//! process). A unique sentinel in the backend error keeps the assertion
+//! robust against any concurrently-running test's log output.
+
+use std::collections::BTreeMap;
+use std::io::Write;
+use std::sync::{Arc, Mutex};
+
+use async_trait::async_trait;
+use picloud_executor_core::{Engine, ExecRequest, InvocationType, Limits};
+use picloud_shared::{
+    AppId, ExecutionId, ModuleScript, ModuleSource, ModuleSourceError, NoopDeadLetterService,
+    NoopDocsService, NoopEventEmitter, NoopHttpService, NoopKvService, RequestId, ScriptId,
+    ScriptSandbox, SdkCallCx, Services,
+};
+use serde_json::Value;
+use tracing_subscriber::fmt::MakeWriter;
+
+const SENTINEL: &str = "connection refused PICLOUD-SENTINEL-9f3a";
+
+struct FailingSource;
+
+#[async_trait]
+impl ModuleSource for FailingSource {
+    async fn lookup(
+        &self,
+        _cx: &SdkCallCx,
+        _name: &str,
+    ) -> Result<Option<ModuleScript>, ModuleSourceError> {
+        Err(ModuleSourceError::Backend(SENTINEL.to_string()))
+    }
+}
+
+/// `MakeWriter` that appends to a shared buffer.
+#[derive(Clone)]
+struct SharedBuf(Arc<Mutex<Vec<u8>>>);
+
+impl Write for SharedBuf {
+    fn write(&mut self, buf: &[u8]) -> std::io::Result<usize> {
+        self.0.lock().unwrap().extend_from_slice(buf);
+        Ok(buf.len())
+    }
+    fn flush(&mut self) -> std::io::Result<()> {
+        Ok(())
+    }
+}
+
+impl<'a> MakeWriter<'a> for SharedBuf {
+    type Writer = SharedBuf;
+    fn make_writer(&'a self) -> Self::Writer {
+        self.clone()
+    }
+}
+
+fn req(app_id: AppId) -> ExecRequest {
+    let execution_id = ExecutionId::new();
+    ExecRequest {
+        execution_id,
+        request_id: RequestId::new(),
+        script_id: ScriptId::new(),
+        script_name: "redaction-test".into(),
+        invocation_type: InvocationType::Http,
+        path: "/x".into(),
+        headers: BTreeMap::new(),
+        body: Value::Null,
+        params: BTreeMap::new(),
+        query: BTreeMap::new(),
+        rest: String::new(),
+        sandbox_overrides: ScriptSandbox::default(),
+        app_id,
+        principal: None,
+        trigger_depth: 0,
+        root_execution_id: execution_id,
+        is_dead_letter_handler: false,
+        event: None,
+    }
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn original_backend_error_is_logged_at_error_level() {
+    let buf = Arc::new(Mutex::new(Vec::<u8>::new()));
+    let subscriber = tracing_subscriber::fmt()
+        .with_writer(SharedBuf(buf.clone()))
+        .with_max_level(tracing::Level::ERROR)
+        .with_ansi(false)
+        .finish();
+    tracing::subscriber::set_global_default(subscriber)
+        .expect("this test owns the global subscriber for its binary");
+
+    let services = Services::new(
+        Arc::new(NoopKvService),
+        Arc::new(NoopDocsService),
+        Arc::new(NoopDeadLetterService),
+        Arc::new(NoopEventEmitter),
+        Arc::new(FailingSource),
+        Arc::new(NoopHttpService),
+        Arc::new(picloud_shared::NoopFilesService),
+        Arc::new(picloud_shared::NoopPubsubService),
+    );
+    let engine = Engine::new(Limits::default(), services);
+
+    let err = engine
+        .execute(r#"import "x" as x; 1"#, req(AppId::new()))
+        .expect_err("backend error should surface");
+
+    // Script-visible: redacted.
+    let msg = format!("{err:?}");
+    assert!(msg.contains("module backend unavailable"), "got {msg}");
+    assert!(
+        !msg.contains("PICLOUD-SENTINEL"),
+        "script error leaked the original: {msg}"
+    );
+
+    // Operator log: the original sentinel IS present, at ERROR level.
+    let logged = String::from_utf8(buf.lock().unwrap().clone()).unwrap();
+    assert!(
+        logged.contains(SENTINEL),
+        "original backend error should be logged; captured: {logged}"
+    );
+    assert!(
+        logged.contains("ERROR"),
+        "should be logged at error level; captured: {logged}"
+    );
+}

crates/executor-core/tests/modules.rs

@@ -0,0 +1,595 @@
+//! v1.1.3 — `PicloudModuleResolver` integration tests.
+#![allow(clippy::needless_raw_string_hashes)] // r#""# is more uniform when many tests embed Rhai sources
+//!
+//! Each test wires an `Engine` with a `CountingModuleSource` (an
+//! in-memory fake), a `Services` bundle, and an `ExecRequest` whose
+//! `app_id` controls the cross-app boundary. The resolver is
+//! exercised end-to-end through `Engine::execute`, so these tests
+//! verify the same code path the `picloud` binary runs at request
+//! time.
+
+use std::collections::{BTreeMap, HashMap};
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::sync::Arc;
+
+use async_trait::async_trait;
+use chrono::{DateTime, Utc};
+use picloud_executor_core::{Engine, ExecRequest, InvocationType, Limits};
+use picloud_shared::{
+    AppId, ExecutionId, ModuleScript, ModuleSource, ModuleSourceError, NoopDeadLetterService,
+    NoopDocsService, NoopEventEmitter, NoopHttpService, NoopKvService, RequestId, ScriptId,
+    ScriptSandbox, SdkCallCx, Services,
+};
+use tokio::sync::Mutex;
+
+/// In-memory `ModuleSource` backed by a `HashMap<(AppId, name)>`.
+/// Tracks total lookup count so tests can assert cache hit/miss.
+#[derive(Default)]
+struct CountingModuleSource {
+    table: Mutex<HashMap<(AppId, String), ModuleScript>>,
+    lookups: AtomicUsize,
+    /// When `Some`, every lookup returns this error instead of the
+    /// table — used by the backend-error test.
+    fail_with: Mutex<Option<String>>,
+}
+
+impl CountingModuleSource {
+    fn new() -> Arc<Self> {
+        Arc::new(Self::default())
+    }
+
+    async fn put(self: &Arc<Self>, app_id: AppId, name: &str, source: &str) -> ScriptId {
+        self.put_with_updated_at(app_id, name, source, Utc::now())
+            .await
+    }
+
+    async fn put_with_updated_at(
+        self: &Arc<Self>,
+        app_id: AppId,
+        name: &str,
+        source: &str,
+        updated_at: DateTime<Utc>,
+    ) -> ScriptId {
+        let script_id = ScriptId::new();
+        self.table.lock().await.insert(
+            (app_id, name.to_string()),
+            ModuleScript {
+                script_id,
+                app_id,
+                name: name.to_string(),
+                source: source.to_string(),
+                updated_at,
+            },
+        );
+        script_id
+    }
+
+    fn lookup_count(&self) -> usize {
+        self.lookups.load(Ordering::SeqCst)
+    }
+}
+
+#[async_trait]
+impl ModuleSource for CountingModuleSource {
+    async fn lookup(
+        &self,
+        cx: &SdkCallCx,
+        name: &str,
+    ) -> Result<Option<ModuleScript>, ModuleSourceError> {
+        self.lookups.fetch_add(1, Ordering::SeqCst);
+        if let Some(err) = self.fail_with.lock().await.as_ref() {
+            return Err(ModuleSourceError::Backend(err.clone()));
+        }
+        Ok(self
+            .table
+            .lock()
+            .await
+            .get(&(cx.app_id, name.to_string()))
+            .cloned())
+    }
+}
+
+fn services_with(modules: Arc<dyn ModuleSource>) -> Services {
+    Services::new(
+        Arc::new(NoopKvService),
+        Arc::new(NoopDocsService),
+        Arc::new(NoopDeadLetterService),
+        Arc::new(NoopEventEmitter),
+        modules,
+        Arc::new(NoopHttpService),
+        Arc::new(picloud_shared::NoopFilesService),
+        Arc::new(picloud_shared::NoopPubsubService),
+    )
+}
+
+fn engine_with(modules: Arc<dyn ModuleSource>) -> Engine {
+    Engine::new(Limits::default(), services_with(modules))
+}
+
+fn req(app_id: AppId) -> ExecRequest {
+    let execution_id = ExecutionId::new();
+    ExecRequest {
+        execution_id,
+        request_id: RequestId::new(),
+        script_id: ScriptId::new(),
+        script_name: "test".into(),
+        invocation_type: InvocationType::Http,
+        path: "/test".into(),
+        headers: BTreeMap::new(),
+        body: serde_json::Value::Null,
+        params: BTreeMap::new(),
+        query: BTreeMap::new(),
+        rest: String::new(),
+        sandbox_overrides: ScriptSandbox::default(),
+        app_id,
+        principal: None,
+        trigger_depth: 0,
+        root_execution_id: execution_id,
+        is_dead_letter_handler: false,
+        event: None,
+    }
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_loads_simple_module() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    source.put(app_id, "math", "fn add(a, b) { a + b }").await;
+
+    let engine = engine_with(source.clone());
+    let resp = engine
+        .execute(r#"import "math" as m; m::add(2, 3)"#, req(app_id))
+        .expect("should execute");
+    assert_eq!(resp.status_code, 200);
+    assert_eq!(resp.body, serde_json::json!(5));
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_cross_app_blocked() {
+    let source = CountingModuleSource::new();
+    let app_a = AppId::new();
+    let app_b = AppId::new();
+    source
+        .put(app_a, "secrets", "fn token() { \"A-token\" }")
+        .await;
+    source
+        .put(app_b, "secrets", "fn token() { \"B-token\" }")
+        .await;
+
+    let engine = engine_with(source.clone());
+
+    // App A sees A's module.
+    let resp = engine
+        .execute(r#"import "secrets" as s; s::token()"#, req(app_a))
+        .unwrap();
+    assert_eq!(resp.body, serde_json::json!("A-token"));
+
+    // App B sees B's module — same name, completely separate value.
+    let resp = engine
+        .execute(r#"import "secrets" as s; s::token()"#, req(app_b))
+        .unwrap();
+    assert_eq!(resp.body, serde_json::json!("B-token"));
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_cross_app_module_not_found() {
+    let source = CountingModuleSource::new();
+    let app_a = AppId::new();
+    let app_b = AppId::new();
+    // Only app A has the module.
+    source.put(app_a, "lonely", "fn ping() { \"pong\" }").await;
+
+    // App B's lookup should return None → resolver surfaces
+    // ErrorModuleNotFound.
+    let engine = engine_with(source.clone());
+    let err = engine
+        .execute(r#"import "lonely" as l; l::ping()"#, req(app_b))
+        .expect_err("cross-app import should fail");
+    let msg = format!("{err:?}");
+    assert!(
+        msg.to_lowercase().contains("module")
+            || msg.to_lowercase().contains("not found")
+            || msg.to_lowercase().contains("lonely"),
+        "expected module-not-found-flavoured error, got {msg}"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_module_not_found() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    let engine = engine_with(source);
+
+    let err = engine
+        .execute(r#"import "doesnotexist" as x; 1"#, req(app_id))
+        .expect_err("unknown module should fail");
+    let msg = format!("{err:?}").to_lowercase();
+    assert!(
+        msg.contains("doesnotexist") || msg.contains("not found"),
+        "expected ErrorModuleNotFound-flavoured error, got {msg}"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_self_import_detected() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    // a imports itself
+    source
+        .put(app_id, "a", r#"import "a" as a; fn nope() { 0 }"#)
+        .await;
+    let engine = engine_with(source);
+
+    let err = engine
+        .execute(r#"import "a" as a; a::nope()"#, req(app_id))
+        .expect_err("self-import should detect cycle");
+    let msg = format!("{err:?}").to_lowercase();
+    assert!(
+        msg.contains("circular") || msg.contains("cycle"),
+        "expected circular-import error, got {msg}"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_circular_detected() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    // a imports b; b imports a; both then declare a fn.
+    source
+        .put(app_id, "a", r#"import "b" as b; fn x() { 0 }"#)
+        .await;
+    source
+        .put(app_id, "b", r#"import "a" as a; fn y() { 0 }"#)
+        .await;
+    let engine = engine_with(source);
+
+    let err = engine
+        .execute(r#"import "a" as a; a::x()"#, req(app_id))
+        .expect_err("circular import should fail");
+    let msg = format!("{err:?}").to_lowercase();
+    assert!(
+        msg.contains("circular") || msg.contains("cycle"),
+        "expected circular-import error, got {msg}"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_depth_limit_enforced() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    // Chain `m0 -> m1 -> ... -> m9` (10 levels). Default depth limit is 8.
+    for i in 0..9 {
+        let next = format!("m{}", i + 1);
+        source
+            .put(
+                app_id,
+                &format!("m{i}"),
+                &format!(r#"import "{next}" as nxt; fn x() {{ 0 }}"#),
+            )
+            .await;
+    }
+    source.put(app_id, "m9", "fn x() { 0 }").await;
+
+    let engine = engine_with(source);
+    let err = engine
+        .execute(r#"import "m0" as m0; m0::x()"#, req(app_id))
+        .expect_err("chain exceeding depth limit should fail");
+    let msg = format!("{err:?}").to_lowercase();
+    assert!(
+        msg.contains("depth"),
+        "expected depth-exceeded error, got {msg}"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_depth_limit_just_under_succeeds() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    // Chain depth 7 (under default 8). m0 -> m1 -> ... -> m6 (terminal).
+    for i in 0..6 {
+        let next = format!("m{}", i + 1);
+        source
+            .put(
+                app_id,
+                &format!("m{i}"),
+                &format!(r#"import "{next}" as nxt; fn x() {{ nxt::x() }}"#),
+            )
+            .await;
+    }
+    source.put(app_id, "m6", "fn x() { 42 }").await;
+
+    let engine = engine_with(source);
+    let resp = engine
+        .execute(r#"import "m0" as m0; m0::x()"#, req(app_id))
+        .expect("chain under depth limit should succeed");
+    assert_eq!(resp.body, serde_json::json!(42));
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_runtime_validation_rejects_top_level_expr() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    // Module has a top-level expression — bypassed the admin gate,
+    // but the resolver re-validates and rejects.
+    source.put(app_id, "bad", r#"42; fn x() { 1 }"#).await;
+    let engine = engine_with(source);
+
+    let err = engine
+        .execute(r#"import "bad" as b; b::x()"#, req(app_id))
+        .expect_err("top-level expr in module should be rejected at resolve");
+    let msg = format!("{err:?}").to_lowercase();
+    assert!(
+        msg.contains("top-level") || msg.contains("module"),
+        "expected module-shape error, got {msg}"
+    );
+}
+
+/// v1.1.4 §10a regression: the backend error must be REDACTED before
+/// it reaches a script. The verbatim message (which can leak internal
+/// infrastructure shape, e.g. "connection refused") must not appear;
+/// the script sees only a stable generic.
+#[tokio::test(flavor = "multi_thread")]
+async fn resolver_backend_error_is_redacted_from_script() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    *source.fail_with.lock().await = Some("connection refused to 10.1.2.3:5432".into());
+    let engine = engine_with(source);
+
+    let err = engine
+        .execute(r#"import "x" as x; 1"#, req(app_id))
+        .expect_err("backend error should propagate");
+    let msg = format!("{err:?}");
+    assert!(
+        msg.contains("module backend unavailable"),
+        "expected redacted generic message, got {msg}"
+    );
+    assert!(
+        !msg.contains("connection refused") && !msg.contains("10.1.2.3"),
+        "redacted message must not leak the backend error, got {msg}"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn module_cache_hit_reuses_compiled_module() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    source.put(app_id, "u", "fn ping() { 1 }").await;
+
+    let engine = engine_with(source.clone());
+
+    // First execution compiles and caches.
+    engine
+        .execute(r#"import "u" as u; u::ping()"#, req(app_id))
+        .unwrap();
+    let lookups_after_first = source.lookup_count();
+    assert_eq!(
+        lookups_after_first, 1,
+        "first invocation should look up once"
+    );
+
+    // Second execution should re-lookup (to compare updated_at) but
+    // serve from cache without recompiling. We can't directly observe
+    // compile-vs-cache here, but we can assert lookup count grew by
+    // one (no spurious extra calls).
+    engine
+        .execute(r#"import "u" as u; u::ping()"#, req(app_id))
+        .unwrap();
+    assert_eq!(source.lookup_count(), 2);
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn module_cache_stale_invalidated_on_updated_at_change() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    let t0 = Utc::now() - chrono::Duration::seconds(10);
+    source
+        .put_with_updated_at(app_id, "u", r#"fn v() { 1 }"#, t0)
+        .await;
+
+    let engine = engine_with(source.clone());
+
+    let resp = engine
+        .execute(r#"import "u" as u; u::v()"#, req(app_id))
+        .unwrap();
+    assert_eq!(resp.body, serde_json::json!(1));
+
+    // Replace with newer updated_at — cache should refresh.
+    let t1 = Utc::now();
+    source
+        .put_with_updated_at(app_id, "u", r#"fn v() { 99 }"#, t1)
+        .await;
+
+    let resp = engine
+        .execute(r#"import "u" as u; u::v()"#, req(app_id))
+        .unwrap();
+    assert_eq!(
+        resp.body,
+        serde_json::json!(99),
+        "edited module should be visible on next invocation"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn module_cache_keyed_by_app() {
+    let source = CountingModuleSource::new();
+    let app_a = AppId::new();
+    let app_b = AppId::new();
+    source.put(app_a, "u", "fn id() { 1 }").await;
+    source.put(app_b, "u", "fn id() { 2 }").await;
+
+    let engine = engine_with(source.clone());
+
+    // Both apps should compile + cache independently; neither sees
+    // the other's compiled module.
+    let resp = engine
+        .execute(r#"import "u" as u; u::id()"#, req(app_a))
+        .unwrap();
+    assert_eq!(resp.body, serde_json::json!(1));
+    let resp = engine
+        .execute(r#"import "u" as u; u::id()"#, req(app_b))
+        .unwrap();
+    assert_eq!(resp.body, serde_json::json!(2));
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn module_cache_lru_evicts_when_capacity_exceeded() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    source.put(app_id, "a", "fn v() { 1 }").await;
+    source.put(app_id, "b", "fn v() { 2 }").await;
+    source.put(app_id, "c", "fn v() { 3 }").await;
+
+    // Capacity 1 — only the most recently used entry stays cached.
+    let engine =
+        Engine::with_module_cache_capacity(Limits::default(), services_with(source.clone()), 1);
+
+    engine
+        .execute(r#"import "a" as m; m::v()"#, req(app_id))
+        .unwrap();
+    engine
+        .execute(r#"import "b" as m; m::v()"#, req(app_id))
+        .unwrap();
+    engine
+        .execute(r#"import "c" as m; m::v()"#, req(app_id))
+        .unwrap();
+
+    // Cache should hold at most one entry.
+    let cache = engine.module_cache().lock().unwrap();
+    assert!(
+        cache.len() <= 1,
+        "cache size {} exceeded capacity 1",
+        cache.len()
+    );
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn endpoint_can_import_module() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    source
+        .put(app_id, "helpers", r#"fn greet(name) { `hello, ${name}` }"#)
+        .await;
+
+    let engine = engine_with(source);
+    let resp = engine
+        .execute(
+            r#"import "helpers" as h; #{ statusCode: 200, body: h::greet("world") }"#,
+            req(app_id),
+        )
+        .unwrap();
+    assert_eq!(resp.status_code, 200);
+    assert_eq!(resp.body, serde_json::json!("hello, world"));
+}
+
+#[tokio::test(flavor = "multi_thread")]
+async fn module_can_import_module() {
+    let source = CountingModuleSource::new();
+    let app_id = AppId::new();
+    source.put(app_id, "inner", "fn three() { 3 }").await;
+    source
+        .put(
+            app_id,
+            "outer",
+            r#"import "inner" as i; fn nine() { i::three() * 3 }"#,
+        )
+        .await;
+    let engine = engine_with(source);
+
+    let resp = engine
+        .execute(r#"import "outer" as o; o::nine()"#, req(app_id))
+        .unwrap();
+    assert_eq!(resp.body, serde_json::json!(9));
+}
+
+#[test]
+fn validate_module_accepts_fn_const_import_only() {
+    let engine = Engine::new(Limits::default(), Services::default());
+    let valid = r#"
+        const PI = 3.14;
+        import "other" as o;
+        fn area(r) { PI * r * r }
+    "#;
+    let v = engine.validate_module(valid).expect("valid module body");
+    assert_eq!(v.imports, vec!["other".to_string()]);
+}
+
+#[test]
+fn validate_module_rejects_top_level_let() {
+    let engine = Engine::new(Limits::default(), Services::default());
+    let bad = "let x = 1; fn f() { x }";
+    let err = engine
+        .validate_module(bad)
+        .expect_err("top-level let should be rejected");
+    let msg = format!("{err:?}").to_lowercase();
+    assert!(msg.contains("top-level") || msg.contains("module"));
+}
+
+#[test]
+fn validate_module_rejects_top_level_expr() {
+    let engine = Engine::new(Limits::default(), Services::default());
+    let bad = "42";
+    let err = engine
+        .validate_module(bad)
+        .expect_err("top-level expr should be rejected");
+    let msg = format!("{err:?}").to_lowercase();
+    assert!(msg.contains("top-level") || msg.contains("module"));
+}
+
+#[test]
+fn validate_module_rejects_top_level_while() {
+    // Avoid `if true { ... }` — Rhai folds constant-condition `if`s
+    // at optimize time, leaving an empty statement list that passes
+    // module-shape validation vacuously. A `while` with a variable
+    // condition isn't folded.
+    let engine = Engine::new(Limits::default(), Services::default());
+    let bad = r#"let i = 0; while i < 1 { i += 1; }"#;
+    let err = engine
+        .validate_module(bad)
+        .expect_err("top-level loop should be rejected");
+    let msg = format!("{err:?}").to_lowercase();
+    assert!(msg.contains("top-level") || msg.contains("module"));
+}
+
+#[test]
+fn validate_endpoint_extracts_literal_imports() {
+    let engine = Engine::new(Limits::default(), Services::default());
+    let src = r#"
+        import "a" as a;
+        import "b" as b;
+        a::run() + b::run()
+    "#;
+    let v = engine
+        .validate(src)
+        .expect("endpoint with imports should parse");
+    assert_eq!(v.imports, vec!["a".to_string(), "b".to_string()]);
+}
+
+#[test]
+fn validate_endpoint_top_level_expr_still_allowed() {
+    // Endpoints can have arbitrary top-level statements — only
+    // modules are restricted. Confirm v1.1.3 didn't tighten endpoints.
+    let engine = Engine::new(Limits::default(), Services::default());
+    let src = r#"let x = 1; #{ statusCode: 200, body: x }"#;
+    engine
+        .validate(src)
+        .expect("endpoints may have top-level statements");
+}
+
+#[test]
+fn validate_endpoint_skips_dynamic_imports_in_imports_list() {
+    // `import some_var as y;` parses but is not a literal-path
+    // import — the dep graph cannot track it. The imports list
+    // should be empty for such a script.
+    let engine = Engine::new(Limits::default(), Services::default());
+    let src = r#"
+        let name = "x";
+        import name as y;
+        y::run()
+    "#;
+    let v = engine.validate(src).expect("dynamic import should parse");
+    assert!(
+        v.imports.is_empty(),
+        "dynamic imports should not appear in the dep-graph imports list, got {:?}",
+        v.imports
+    );
+}

crates/executor-core/tests/sdk_docs.rs

@@ -0,0 +1,524 @@
+//! `docs::` SDK bridge integration tests — runs a real Rhai engine
+//! against an in-memory `DocsService` impl. Mirrors `tests/sdk_kv.rs`:
+//! `tokio::task::spawn_blocking` so the bridge's `block_on` has a
+//! reachable runtime.
+
+use std::collections::{BTreeMap, HashMap};
+use std::sync::Arc;
+
+use async_trait::async_trait;
+use chrono::Utc;
+use picloud_executor_core::{Engine, ExecRequest, InvocationType, Limits};
+use picloud_shared::{
+    AppId, DocId, DocRow, DocsError, DocsListPage, DocsService, ExecutionId, NoopDeadLetterService,
+    NoopEventEmitter, NoopHttpService, NoopKvService, NoopModuleSource, RequestId, ScriptId,
+    ScriptSandbox, SdkCallCx, Services,
+};
+use serde_json::{json, Value};
+use tokio::sync::Mutex;
+use uuid::Uuid;
+
+#[derive(Default)]
+struct InMemoryDocs {
+    data: Mutex<HashMap<(AppId, String, DocId), DocRow>>,
+}
+
+#[async_trait]
+impl DocsService for InMemoryDocs {
+    async fn create(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        data: Value,
+    ) -> Result<DocId, DocsError> {
+        if !data.is_object() {
+            return Err(DocsError::InvalidData);
+        }
+        let id = Uuid::new_v4();
+        let now = Utc::now();
+        let row = DocRow {
+            id,
+            data,
+            created_at: now,
+            updated_at: now,
+        };
+        self.data
+            .lock()
+            .await
+            .insert((cx.app_id, collection.to_string(), id), row);
+        Ok(id)
+    }
+
+    async fn get(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: DocId,
+    ) -> Result<Option<DocRow>, DocsError> {
+        Ok(self
+            .data
+            .lock()
+            .await
+            .get(&(cx.app_id, collection.to_string(), id))
+            .cloned())
+    }
+
+    async fn find(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        filter: Value,
+    ) -> Result<Vec<DocRow>, DocsError> {
+        // Tiny eval: extract top-level equalities + $in arrays + $gt
+        // (text lex) so the bridge tests can run end-to-end against a
+        // fake. This fake mirrors the real service's reject-unsupported
+        // contract so the v1.2-pointer-error test goes through the
+        // bridge's error-propagation path.
+        let map = self.data.lock().await;
+        let obj = filter
+            .as_object()
+            .ok_or_else(|| DocsError::InvalidFilter("filter must be a map/object".into()))?;
+        reject_unsupported_operators(obj)?;
+        let mut out: Vec<DocRow> = map
+            .iter()
+            .filter(|((a, c, _), _)| *a == cx.app_id && c == collection)
+            .map(|(_, v)| v.clone())
+            .filter(|row| matches_simple(&row.data, obj))
+            .collect();
+        if let Some(limit) = obj.get("$limit").and_then(Value::as_u64) {
+            out.truncate(usize::try_from(limit).unwrap_or(usize::MAX));
+        }
+        Ok(out)
+    }
+
+    async fn find_one(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        filter: Value,
+    ) -> Result<Option<DocRow>, DocsError> {
+        Ok(self.find(cx, collection, filter).await?.into_iter().next())
+    }
+
+    async fn update(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: DocId,
+        data: Value,
+    ) -> Result<(), DocsError> {
+        if !data.is_object() {
+            return Err(DocsError::InvalidData);
+        }
+        let mut map = self.data.lock().await;
+        let key = (cx.app_id, collection.to_string(), id);
+        let Some(row) = map.get_mut(&key) else {
+            return Err(DocsError::NotFound);
+        };
+        row.data = data;
+        row.updated_at = Utc::now();
+        Ok(())
+    }
+
+    async fn delete(&self, cx: &SdkCallCx, collection: &str, id: DocId) -> Result<bool, DocsError> {
+        Ok(self
+            .data
+            .lock()
+            .await
+            .remove(&(cx.app_id, collection.to_string(), id))
+            .is_some())
+    }
+
+    async fn list(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        _cursor: Option<&str>,
+        _limit: u32,
+    ) -> Result<DocsListPage, DocsError> {
+        let mut docs: Vec<DocRow> = self
+            .data
+            .lock()
+            .await
+            .iter()
+            .filter(|((a, c, _), _)| *a == cx.app_id && c == collection)
+            .map(|(_, v)| v.clone())
+            .collect();
+        docs.sort_by_key(|d| d.id);
+        Ok(DocsListPage {
+            docs,
+            next_cursor: None,
+        })
+    }
+}
+
+/// Scan an operator object for any `$xxx` key not in the v1.1.2
+/// allowlist and return the same shape of error the real parser
+/// emits. Top-level `$limit` is the only allowed modifier the fake
+/// engages with; the unsupported test passes `$regex`.
+fn reject_unsupported_operators(obj: &serde_json::Map<String, Value>) -> Result<(), DocsError> {
+    const SUPPORTED_TOP_LEVEL: &[&str] = &["$limit", "$sort"];
+    const SUPPORTED_NESTED: &[&str] = &["$eq", "$ne", "$gt", "$gte", "$lt", "$lte", "$in"];
+    for (key, value) in obj {
+        if let Some(stripped) = key.strip_prefix('$') {
+            if !SUPPORTED_TOP_LEVEL.contains(&key.as_str()) {
+                return Err(DocsError::UnsupportedOperator(format!(
+                    "docs::find: top-level modifier '${stripped}' is not supported in v1.1.2; planned for v1.2 advanced query"
+                )));
+            }
+            continue;
+        }
+        if let Some(inner) = value.as_object() {
+            for op_key in inner.keys() {
+                if op_key.starts_with('$') && !SUPPORTED_NESTED.contains(&op_key.as_str()) {
+                    return Err(DocsError::UnsupportedOperator(format!(
+                        "docs::find: operator '{op_key}' is not supported in v1.1.2; planned for v1.2 advanced query"
+                    )));
+                }
+            }
+        }
+    }
+    Ok(())
+}
+
+fn matches_simple(data: &Value, filter: &serde_json::Map<String, Value>) -> bool {
+    for (key, want) in filter {
+        if key.starts_with('$') {
+            // $limit handled in the find body.
+            continue;
+        }
+        let actual = data.get(key);
+        if let Some(obj) = want.as_object() {
+            // operator object — handle $in and $gt only (enough for
+            // the bridge tests to exercise the round-trip).
+            if let Some(arr) = obj.get("$in").and_then(Value::as_array) {
+                let Some(actual) = actual else {
+                    return false;
+                };
+                if !arr.iter().any(|v| v == actual) {
+                    return false;
+                }
+                continue;
+            }
+            if let Some(gt) = obj.get("$gt") {
+                let Some(actual) = actual else {
+                    return false;
+                };
+                let a = actual.as_str().unwrap_or("");
+                let b = gt.as_str().unwrap_or("");
+                if a <= b {
+                    return false;
+                }
+                continue;
+            }
+            return false;
+        }
+        if Some(want) != actual {
+            return false;
+        }
+    }
+    true
+}
+
+fn make_engine() -> Arc<Engine> {
+    let services = Services::new(
+        Arc::new(NoopKvService),
+        Arc::new(InMemoryDocs::default()),
+        Arc::new(NoopDeadLetterService),
+        Arc::new(NoopEventEmitter),
+        Arc::new(NoopModuleSource),
+        Arc::new(NoopHttpService),
+        Arc::new(picloud_shared::NoopFilesService),
+        Arc::new(picloud_shared::NoopPubsubService),
+    );
+    Arc::new(Engine::new(Limits::default(), services))
+}
+
+fn baseline_request(app_id: AppId) -> ExecRequest {
+    let execution_id = ExecutionId::new();
+    ExecRequest {
+        execution_id,
+        request_id: RequestId::new(),
+        script_id: ScriptId::new(),
+        script_name: "docs-test".into(),
+        invocation_type: InvocationType::Http,
+        path: "/docs-test".into(),
+        headers: BTreeMap::new(),
+        body: Value::Null,
+        params: BTreeMap::new(),
+        query: BTreeMap::new(),
+        rest: String::new(),
+        sandbox_overrides: ScriptSandbox::default(),
+        app_id,
+        principal: None,
+        trigger_depth: 0,
+        root_execution_id: execution_id,
+        is_dead_letter_handler: false,
+        event: None,
+    }
+}
+
+async fn run_script(engine: Arc<Engine>, src: &str, req: ExecRequest) -> Value {
+    let src = src.to_string();
+    tokio::task::spawn_blocking(move || engine.execute(&src, req))
+        .await
+        .expect("spawn_blocking should not panic")
+        .expect("script execution should succeed")
+        .body
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_create_then_get_round_trip() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let users = docs::collection("users");
+        let id = users.create(#{ name: "Alice", tier: "gold" });
+        let doc = users.get(id);
+        #{ id_matches: doc.id == id, data_name: doc.data.name }
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    let obj = body.as_object().unwrap();
+    assert_eq!(obj["id_matches"], json!(true));
+    assert_eq!(obj["data_name"], json!("Alice"));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_get_missing_returns_unit() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        let v = c.get("00000000-0000-0000-0000-000000000000");
+        v == ()
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!(true));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_get_with_invalid_uuid_throws() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"docs::collection("users").get("not-a-uuid")"#;
+    let req = baseline_request(app);
+    let err = tokio::task::spawn_blocking(move || engine.execute(src, req))
+        .await
+        .unwrap()
+        .expect_err("invalid uuid should throw");
+    assert!(format!("{err:?}").contains("invalid id"));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_find_equality_returns_matches() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        c.create(#{ tier: "gold" });
+        c.create(#{ tier: "silver" });
+        c.create(#{ tier: "gold" });
+        let golds = c.find(#{ tier: "gold" });
+        golds.len()
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!(2));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_find_with_in_operator() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        c.create(#{ tier: "gold" });
+        c.create(#{ tier: "silver" });
+        c.create(#{ tier: "platinum" });
+        let hits = c.find(#{ tier: #{ "$in": ["gold", "platinum"] } });
+        hits.len()
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!(2));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_find_with_gt_comparison() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("events");
+        c.create(#{ when: "2026-01-15" });
+        c.create(#{ when: "2026-03-15" });
+        c.create(#{ when: "2026-05-15" });
+        let recent = c.find(#{ when: #{ "$gt": "2026-02-01" } });
+        recent.len()
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!(2));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_find_one_returns_envelope_or_unit() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        c.create(#{ tier: "gold" });
+        let hit = c.find_one(#{ tier: "gold" });
+        let miss = c.find_one(#{ tier: "platinum" });
+        #{ hit_has_data: hit.data.tier == "gold", miss_is_unit: miss == () }
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    let obj = body.as_object().unwrap();
+    assert_eq!(obj["hit_has_data"], json!(true));
+    assert_eq!(obj["miss_is_unit"], json!(true));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_update_then_get_reflects_change() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        let id = c.create(#{ name: "Alice", tier: "gold" });
+        c.update(id, #{ name: "Alice", tier: "platinum" });
+        c.get(id).data.tier
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!("platinum"));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_update_missing_throws() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        c.update("00000000-0000-0000-0000-000000000000", #{ x: 1 })
+    "#;
+    let req = baseline_request(app);
+    let err = tokio::task::spawn_blocking(move || engine.execute(src, req))
+        .await
+        .unwrap()
+        .expect_err("update missing should throw");
+    assert!(format!("{err:?}").contains("not found"));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_delete_returns_was_present() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        let nope = c.delete("00000000-0000-0000-0000-000000000000");
+        let id = c.create(#{ x: 1 });
+        let yep = c.delete(id);
+        #{ nope: nope, yep: yep }
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!({ "nope": false, "yep": true }));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_unsupported_operator_throws_with_v1_2_pointer() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        c.find(#{ name: #{ "$regex": "^A" } })
+    "#;
+    let req = baseline_request(app);
+    let err = tokio::task::spawn_blocking(move || engine.execute(src, req))
+        .await
+        .unwrap()
+        .expect_err("unsupported operator should throw");
+    let msg = format!("{err:?}");
+    assert!(msg.contains("$regex"), "msg: {msg}");
+    assert!(msg.contains("v1.2"), "msg: {msg}");
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_empty_collection_name_throws() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"docs::collection("")"#;
+    let req = baseline_request(app);
+    let err = tokio::task::spawn_blocking(move || engine.execute(src, req))
+        .await
+        .unwrap()
+        .expect_err("empty collection should throw");
+    assert!(format!("{err:?}").contains("docs::collection"));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_list_returns_docs_array() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        c.create(#{ a: 1 });
+        c.create(#{ a: 2 });
+        let page = c.list();
+        page.docs.len()
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!(2));
+}
+
+/// Cross-app isolation through the bridge — script with `app_id = A`
+/// must NOT see documents written from `app_id = B` even when the
+/// (collection, id) tuple is shared. The bridge captures `cx.app_id`
+/// via `Arc<SdkCallCx>` and the service derives storage `app_id` from
+/// it (never from a script arg).
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_bridge_preserves_cross_app_isolation() {
+    let engine = make_engine();
+    let app_a = AppId::new();
+    let app_b = AppId::new();
+
+    let writer = r#"
+        let c = docs::collection("shared");
+        let id = c.create(#{ from: "a" });
+        id
+    "#;
+    let id_a = run_script(engine.clone(), writer, baseline_request(app_a)).await;
+    let id_a_str = id_a.as_str().unwrap().to_string();
+
+    // App B looks up the same id under the same collection — should
+    // see nothing because the service keyed it by app_id = A.
+    let reader_src = format!(
+        r#"
+        let c = docs::collection("shared");
+        let v = c.get("{id_a_str}");
+        v == ()
+    "#
+    );
+    let body = run_script(engine, &reader_src, baseline_request(app_b)).await;
+    assert_eq!(body, json!(true));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn docs_envelope_has_id_data_created_at_updated_at() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = docs::collection("users");
+        let id = c.create(#{ name: "Alice" });
+        let doc = c.get(id);
+        // Probe each envelope field is present + correctly typed.
+        #{
+            has_id:         type_of(doc.id) == "string",
+            has_data:       type_of(doc.data) == "map",
+            has_created_at: type_of(doc.created_at) == "string",
+            has_updated_at: type_of(doc.updated_at) == "string",
+            user_field:     doc.data.name
+        }
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    let obj = body.as_object().unwrap();
+    assert_eq!(obj["has_id"], json!(true));
+    assert_eq!(obj["has_data"], json!(true));
+    assert_eq!(obj["has_created_at"], json!(true));
+    assert_eq!(obj["has_updated_at"], json!(true));
+    assert_eq!(obj["user_field"], json!("Alice"));
+}

crates/executor-core/tests/sdk_files.rs

@@ -0,0 +1,334 @@
+//! `files::` SDK bridge integration tests — runs a real Rhai engine
+//! against an in-memory `FilesService` impl. Mirrors `tests/sdk_kv.rs`:
+//! `tokio::task::spawn_blocking` so the bridge's `block_on` has a
+//! reachable runtime. Exercises the actual Rhai surface — blob in/out,
+//! the metadata map shape, and the missing-required-field throw.
+
+use std::collections::BTreeMap;
+use std::sync::Arc;
+
+use async_trait::async_trait;
+use picloud_executor_core::{Engine, ExecRequest, InvocationType, Limits};
+use picloud_shared::{
+    AppId, ExecutionId, FileMeta, FileUpdate, FilesError, FilesListPage, FilesService, NewFile,
+    NoopDeadLetterService, NoopDocsService, NoopEventEmitter, NoopHttpService, NoopKvService,
+    NoopModuleSource, RequestId, ScriptId, ScriptSandbox, SdkCallCx, Services,
+};
+use serde_json::{json, Value};
+use tokio::sync::Mutex;
+use uuid::Uuid;
+
+#[derive(Default)]
+struct InMemoryFiles {
+    #[allow(clippy::type_complexity)]
+    data: Mutex<BTreeMap<(AppId, String, Uuid), (FileMeta, Vec<u8>)>>,
+}
+
+/// The in-memory fake doesn't exercise the real checksum path (the
+/// `FsFilesRepo` tempdir tests in manager-core cover SHA-256); a stable
+/// placeholder keeps the metadata map non-empty.
+fn fake_checksum(bytes: &[u8]) -> String {
+    format!("len-{}", bytes.len())
+}
+
+#[async_trait]
+impl FilesService for InMemoryFiles {
+    async fn create(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        new: NewFile,
+    ) -> Result<Uuid, FilesError> {
+        if collection.is_empty() {
+            return Err(FilesError::InvalidCollection("empty".into()));
+        }
+        new.validate(100 * 1024 * 1024)?;
+        let id = Uuid::new_v4();
+        let now = chrono::Utc::now();
+        let meta = FileMeta {
+            id,
+            collection: collection.to_string(),
+            name: new.name.clone(),
+            content_type: new.content_type.clone(),
+            size: new.data.len() as u64,
+            checksum: fake_checksum(&new.data),
+            created_at: now,
+            updated_at: now,
+        };
+        self.data
+            .lock()
+            .await
+            .insert((cx.app_id, collection.to_string(), id), (meta, new.data));
+        Ok(id)
+    }
+
+    async fn head(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: &str,
+    ) -> Result<Option<FileMeta>, FilesError> {
+        let Ok(uuid) = Uuid::parse_str(id) else {
+            return Ok(None);
+        };
+        Ok(self
+            .data
+            .lock()
+            .await
+            .get(&(cx.app_id, collection.to_string(), uuid))
+            .map(|(m, _)| m.clone()))
+    }
+
+    async fn get(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: &str,
+    ) -> Result<Option<Vec<u8>>, FilesError> {
+        let Ok(uuid) = Uuid::parse_str(id) else {
+            return Ok(None);
+        };
+        Ok(self
+            .data
+            .lock()
+            .await
+            .get(&(cx.app_id, collection.to_string(), uuid))
+            .map(|(_, b)| b.clone()))
+    }
+
+    async fn update(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: &str,
+        upd: FileUpdate,
+    ) -> Result<(), FilesError> {
+        upd.validate(100 * 1024 * 1024)?;
+        let Ok(uuid) = Uuid::parse_str(id) else {
+            return Err(FilesError::NotFound);
+        };
+        let mut data = self.data.lock().await;
+        let key = (cx.app_id, collection.to_string(), uuid);
+        let Some((meta, _)) = data.get(&key).cloned() else {
+            return Err(FilesError::NotFound);
+        };
+        let mut meta = meta;
+        if let Some(n) = upd.name {
+            meta.name = n;
+        }
+        if let Some(ct) = upd.content_type {
+            meta.content_type = ct;
+        }
+        meta.size = upd.data.len() as u64;
+        meta.checksum = fake_checksum(&upd.data);
+        data.insert(key, (meta, upd.data));
+        Ok(())
+    }
+
+    async fn delete(&self, cx: &SdkCallCx, collection: &str, id: &str) -> Result<bool, FilesError> {
+        let Ok(uuid) = Uuid::parse_str(id) else {
+            return Ok(false);
+        };
+        Ok(self
+            .data
+            .lock()
+            .await
+            .remove(&(cx.app_id, collection.to_string(), uuid))
+            .is_some())
+    }
+
+    async fn list(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        _cursor: Option<&str>,
+        _limit: u32,
+    ) -> Result<FilesListPage, FilesError> {
+        let data = self.data.lock().await;
+        let files: Vec<FileMeta> = data
+            .iter()
+            .filter(|((a, c, _), _)| *a == cx.app_id && c == collection)
+            .map(|(_, (m, _))| m.clone())
+            .collect();
+        Ok(FilesListPage {
+            files,
+            next_cursor: None,
+        })
+    }
+}
+
+fn make_engine() -> Arc<Engine> {
+    let services = Services::new(
+        Arc::new(NoopKvService),
+        Arc::new(NoopDocsService),
+        Arc::new(NoopDeadLetterService),
+        Arc::new(NoopEventEmitter),
+        Arc::new(NoopModuleSource),
+        Arc::new(NoopHttpService),
+        Arc::new(InMemoryFiles::default()),
+        Arc::new(picloud_shared::NoopPubsubService),
+    );
+    Arc::new(Engine::new(Limits::default(), services))
+}
+
+fn baseline_request(app_id: AppId) -> ExecRequest {
+    let execution_id = ExecutionId::new();
+    ExecRequest {
+        execution_id,
+        request_id: RequestId::new(),
+        script_id: ScriptId::new(),
+        script_name: "files-test".into(),
+        invocation_type: InvocationType::Http,
+        path: "/files-test".into(),
+        headers: BTreeMap::new(),
+        body: Value::Null,
+        params: BTreeMap::new(),
+        query: BTreeMap::new(),
+        rest: String::new(),
+        sandbox_overrides: ScriptSandbox::default(),
+        app_id,
+        principal: None,
+        trigger_depth: 0,
+        root_execution_id: execution_id,
+        is_dead_letter_handler: false,
+        event: None,
+    }
+}
+
+async fn run_script(engine: Arc<Engine>, src: &str, req: ExecRequest) -> Value {
+    let src = src.to_string();
+    tokio::task::spawn_blocking(move || engine.execute(&src, req))
+        .await
+        .expect("spawn_blocking should not panic")
+        .expect("script execution should succeed")
+        .body
+}
+
+async fn run_script_err(engine: Arc<Engine>, src: &str, req: ExecRequest) -> String {
+    let src = src.to_string();
+    let res = tokio::task::spawn_blocking(move || engine.execute(&src, req))
+        .await
+        .expect("spawn_blocking should not panic");
+    format!("{:?}", res.expect_err("script should error"))
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn files_create_get_round_trip_via_blob() {
+    let engine = make_engine();
+    let app = AppId::new();
+    // base64("hello") = "aGVsbG8="; decode → blob; create; get back; encode.
+    let src = r#"
+        let c = files::collection("avatars");
+        let data = base64::decode("aGVsbG8=");
+        let id = c.create(#{ name: "a.txt", content_type: "text/plain", data: data });
+        let back = c.get(id);
+        base64::encode(back)
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!("aGVsbG8="));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn files_head_returns_metadata_map() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = files::collection("avatars");
+        let data = base64::decode("aGVsbG8=");
+        let id = c.create(#{ name: "a.txt", content_type: "text/plain", data: data });
+        let meta = c.head(id);
+        #{ name: meta.name, content_type: meta.content_type, size: meta.size, has_checksum: meta.checksum != () }
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(
+        body,
+        json!({ "name": "a.txt", "content_type": "text/plain", "size": 5, "has_checksum": true })
+    );
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn files_get_and_head_missing_return_unit() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = files::collection("avatars");
+        let g = c.get("00000000-0000-0000-0000-000000000000");
+        let h = c.head("00000000-0000-0000-0000-000000000000");
+        #{ g: g == (), h: h == () }
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!({ "g": true, "h": true }));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn files_update_then_delete() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = files::collection("avatars");
+        let id = c.create(#{ name: "a", content_type: "text/plain", data: base64::decode("YQ==") });
+        c.update(id, #{ data: base64::decode("YmM=") });   // "bc"
+        let after = base64::encode(c.get(id));
+        let removed = c.delete(id);
+        let gone = c.delete(id);
+        #{ after: after, removed: removed, gone: gone }
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(
+        body,
+        json!({ "after": "YmM=", "removed": true, "gone": false })
+    );
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn files_create_missing_data_throws_naming_field() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = files::collection("avatars");
+        c.create(#{ name: "a", content_type: "text/plain" })
+    "#;
+    let err = run_script_err(engine, src, baseline_request(app)).await;
+    assert!(
+        err.contains("data"),
+        "error should name the missing field: {err}"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn files_create_missing_name_throws_naming_field() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = files::collection("avatars");
+        c.create(#{ content_type: "text/plain", data: base64::decode("YQ==") })
+    "#;
+    let err = run_script_err(engine, src, baseline_request(app)).await;
+    assert!(
+        err.contains("name"),
+        "error should name the missing field: {err}"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn files_empty_collection_name_throws() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let err = run_script_err(engine, r#"files::collection("")"#, baseline_request(app)).await;
+    assert!(err.to_lowercase().contains("empty"), "got {err}");
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn files_list_returns_files_array() {
+    let engine = make_engine();
+    let app = AppId::new();
+    let src = r#"
+        let c = files::collection("avatars");
+        c.create(#{ name: "a", content_type: "text/plain", data: base64::decode("YQ==") });
+        c.create(#{ name: "b", content_type: "text/plain", data: base64::decode("Yg==") });
+        let page = c.list();
+        page.files.len()
+    "#;
+    let body = run_script(engine, src, baseline_request(app)).await;
+    assert_eq!(body, json!(2));
+}

crates/executor-core/tests/sdk_http.rs

@@ -0,0 +1,336 @@
+//! Bridge integration for the `http::*` SDK (v1.1.4).
+//!
+//! Runs a real Rhai engine under `spawn_blocking` against an in-memory
+//! `HttpService` fake that records the last request and returns a
+//! configured response (or error). This exercises the full bridge:
+//! option parsing, body dispatch, response→map projection, the
+//! throw-on-network-error / no-throw-on-non-2xx convention, and that
+//! `cx.app_id` / `cx.script_id` are forwarded for attribution.
+
+use std::collections::BTreeMap;
+use std::sync::{Arc, Mutex};
+
+use async_trait::async_trait;
+use picloud_executor_core::{Engine, ExecRequest, InvocationType, Limits};
+use picloud_shared::{
+    AppId, ExecutionId, HttpError, HttpRequest, HttpResponse, HttpService, NoopDeadLetterService,
+    NoopDocsService, NoopEventEmitter, NoopKvService, NoopModuleSource, RequestId, ScriptId,
+    ScriptSandbox, Services,
+};
+use serde_json::{json, Value};
+
+/// What the fake returns. Either a canned response or an error.
+#[derive(Clone)]
+enum Behavior {
+    Respond(HttpResponse),
+    Fail(String), // becomes HttpError::Network
+}
+
+#[derive(Default)]
+struct Recorded {
+    last: Option<HttpRequest>,
+    last_app: Option<AppId>,
+    last_script: Option<String>,
+}
+
+struct FakeHttp {
+    behavior: Behavior,
+    recorded: Mutex<Recorded>,
+}
+
+impl FakeHttp {
+    fn responding(status: u16, content_type: &str, body: &str) -> Arc<Self> {
+        let mut headers = BTreeMap::new();
+        headers.insert("content-type".into(), content_type.into());
+        Arc::new(Self {
+            behavior: Behavior::Respond(HttpResponse {
+                status,
+                headers,
+                body_raw: body.into(),
+            }),
+            recorded: Mutex::new(Recorded::default()),
+        })
+    }
+
+    fn failing(msg: &str) -> Arc<Self> {
+        Arc::new(Self {
+            behavior: Behavior::Fail(msg.into()),
+            recorded: Mutex::new(Recorded::default()),
+        })
+    }
+}
+
+#[async_trait]
+impl HttpService for FakeHttp {
+    async fn request(
+        &self,
+        cx: &picloud_shared::SdkCallCx,
+        req: HttpRequest,
+    ) -> Result<HttpResponse, HttpError> {
+        {
+            let mut r = self.recorded.lock().unwrap();
+            r.last = Some(req.clone());
+            r.last_app = Some(cx.app_id);
+            r.last_script = Some(cx.script_id.to_string());
+        }
+        match &self.behavior {
+            Behavior::Respond(resp) => Ok(resp.clone()),
+            Behavior::Fail(msg) => Err(HttpError::Network(msg.clone())),
+        }
+    }
+}
+
+fn engine_with(http: Arc<dyn HttpService>) -> Arc<Engine> {
+    let services = Services::new(
+        Arc::new(NoopKvService),
+        Arc::new(NoopDocsService),
+        Arc::new(NoopDeadLetterService),
+        Arc::new(NoopEventEmitter),
+        Arc::new(NoopModuleSource),
+        http,
+        Arc::new(picloud_shared::NoopFilesService),
+        Arc::new(picloud_shared::NoopPubsubService),
+    );
+    Arc::new(Engine::new(Limits::default(), services))
+}
+
+fn baseline_request(app_id: AppId, script_id: ScriptId) -> ExecRequest {
+    let execution_id = ExecutionId::new();
+    ExecRequest {
+        execution_id,
+        request_id: RequestId::new(),
+        script_id,
+        script_name: "http-test".into(),
+        invocation_type: InvocationType::Http,
+        path: "/http-test".into(),
+        headers: BTreeMap::new(),
+        body: Value::Null,
+        params: BTreeMap::new(),
+        query: BTreeMap::new(),
+        rest: String::new(),
+        sandbox_overrides: ScriptSandbox::default(),
+        app_id,
+        principal: None,
+        trigger_depth: 0,
+        root_execution_id: execution_id,
+        is_dead_letter_handler: false,
+        event: None,
+    }
+}
+
+async fn run(engine: Arc<Engine>, src: &str, req: ExecRequest) -> Value {
+    let src = src.to_string();
+    tokio::task::spawn_blocking(move || engine.execute(&src, req))
+        .await
+        .expect("spawn_blocking should not panic")
+        .expect("script execution should succeed")
+        .body
+}
+
+async fn run_err(engine: Arc<Engine>, src: &str, req: ExecRequest) -> String {
+    let src = src.to_string();
+    let err = tokio::task::spawn_blocking(move || engine.execute(&src, req))
+        .await
+        .unwrap()
+        .expect_err("script should throw");
+    format!("{err:?}")
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn get_returns_status_and_json_body() {
+    let http = FakeHttp::responding(200, "application/json", r#"{"ok":true,"n":7}"#);
+    let engine = engine_with(http.clone());
+    let src = r#"
+        let r = http::get("https://api.example.com/x");
+        #{ status: r.status, ok: r.body.ok, n: r.body.n }
+    "#;
+    let body = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    assert_eq!(body, json!({ "status": 200, "ok": true, "n": 7 }));
+    // GET carries no body.
+    assert!(http
+        .recorded
+        .lock()
+        .unwrap()
+        .last
+        .as_ref()
+        .unwrap()
+        .body
+        .is_none());
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn non_json_body_stays_string() {
+    let http = FakeHttp::responding(200, "text/plain", "plain text");
+    let engine = engine_with(http);
+    let src = r#"http::get("https://x/").body"#;
+    let body = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    assert_eq!(body, json!("plain text"));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn empty_body_is_unit() {
+    let http = FakeHttp::responding(204, "text/plain", "");
+    let engine = engine_with(http);
+    let src = r#"
+        let r = http::get("https://x/");
+        #{ is_unit: r.body == (), raw: r.body_raw }
+    "#;
+    let body = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    assert_eq!(body, json!({ "is_unit": true, "raw": "" }));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn post_map_body_is_json_encoded() {
+    let http = FakeHttp::responding(200, "application/json", "{}");
+    let engine = engine_with(http.clone());
+    let src = r#"http::post("https://hooks/x", #{ text: "hello", n: 3 }).status"#;
+    let _ = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    let rec = http.recorded.lock().unwrap();
+    let req = rec.last.as_ref().unwrap();
+    assert_eq!(req.method, "POST");
+    assert_eq!(req.content_type.as_deref(), Some("application/json"));
+    let sent: Value = serde_json::from_slice(req.body.as_ref().unwrap()).unwrap();
+    assert_eq!(sent, json!({ "text": "hello", "n": 3 }));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn post_string_body_is_text_plain() {
+    let http = FakeHttp::responding(200, "text/plain", "ok");
+    let engine = engine_with(http.clone());
+    let src = r#"http::post("https://x/", "raw payload").status"#;
+    let _ = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    let rec = http.recorded.lock().unwrap();
+    let req = rec.last.as_ref().unwrap();
+    assert_eq!(req.content_type.as_deref(), Some("text/plain"));
+    assert_eq!(req.body.as_deref(), Some(&b"raw payload"[..]));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn post_unit_body_sends_nothing() {
+    let http = FakeHttp::responding(200, "text/plain", "ok");
+    let engine = engine_with(http.clone());
+    let src = r#"http::post("https://x/", ()).status"#;
+    let _ = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    assert!(http
+        .recorded
+        .lock()
+        .unwrap()
+        .last
+        .as_ref()
+        .unwrap()
+        .body
+        .is_none());
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn custom_headers_and_timeout_forwarded() {
+    let http = FakeHttp::responding(200, "text/plain", "ok");
+    let engine = engine_with(http.clone());
+    let src = r#"
+        http::get("https://x/", #{
+            headers: #{ "Authorization": "Bearer t0ken" },
+            timeout_ms: 4200,
+        }).status
+    "#;
+    let _ = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    let rec = http.recorded.lock().unwrap();
+    let req = rec.last.as_ref().unwrap();
+    assert_eq!(
+        req.headers.get("Authorization").map(String::as_str),
+        Some("Bearer t0ken")
+    );
+    assert_eq!(req.timeout_ms, 4200);
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn unknown_option_key_throws() {
+    let http = FakeHttp::responding(200, "text/plain", "ok");
+    let engine = engine_with(http);
+    let src = r#"http::get("https://x/", #{ timeoutms: 1000 })"#; // typo
+    let err = run_err(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    assert!(err.contains("unknown option key"), "got {err}");
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn timeout_above_max_throws() {
+    let http = FakeHttp::responding(200, "text/plain", "ok");
+    let engine = engine_with(http);
+    let src = r#"http::get("https://x/", #{ timeout_ms: 99999 })"#;
+    let err = run_err(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    assert!(err.contains("maximum"), "got {err}");
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn non_2xx_does_not_throw() {
+    let http = FakeHttp::responding(503, "text/plain", "down");
+    let engine = engine_with(http);
+    let src = r#"http::get("https://x/").status"#;
+    let body = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    assert_eq!(body, json!(503));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn network_error_throws_with_http_prefix() {
+    let http = FakeHttp::failing("connection refused");
+    let engine = engine_with(http);
+    let src = r#"http::get("https://x/")"#;
+    let err = run_err(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    assert!(err.contains("http:"), "expected http: prefix, got {err}");
+    assert!(err.contains("connection refused"), "got {err}");
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn post_form_url_encodes() {
+    let http = FakeHttp::responding(200, "text/plain", "ok");
+    let engine = engine_with(http.clone());
+    let src = r#"http::post_form("https://x/login", #{ user: "alice", pw: "p@ss word" }).status"#;
+    let _ = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    let rec = http.recorded.lock().unwrap();
+    let req = rec.last.as_ref().unwrap();
+    assert_eq!(
+        req.content_type.as_deref(),
+        Some("application/x-www-form-urlencoded")
+    );
+    let body = String::from_utf8(req.body.clone().unwrap()).unwrap();
+    // order is map iteration order; assert both pairs present, encoded.
+    assert!(body.contains("user=alice"), "got {body}");
+    assert!(body.contains("pw=p%40ss+word"), "got {body}");
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn request_escape_hatch_arbitrary_method() {
+    let http = FakeHttp::responding(200, "text/plain", "ok");
+    let engine = engine_with(http.clone());
+    let src = r#"http::request("OPTIONS", "https://x/").status"#;
+    let _ = run(engine, src, baseline_request(AppId::new(), ScriptId::new())).await;
+    assert_eq!(
+        http.recorded.lock().unwrap().last.as_ref().unwrap().method,
+        "OPTIONS"
+    );
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn default_user_agent_carries_script_id() {
+    let http = FakeHttp::responding(200, "text/plain", "ok");
+    let engine = engine_with(http.clone());
+    let script_id = ScriptId::new();
+    let src = r#"http::get("https://x/").status"#;
+    let _ = run(engine, src, baseline_request(AppId::new(), script_id)).await;
+    let rec = http.recorded.lock().unwrap();
+    // The bridge forwards script_id on the request; the manager-core
+    // impl turns it into the User-Agent. Here we assert the forward.
+    assert_eq!(
+        rec.last.as_ref().unwrap().script_id.as_deref(),
+        Some(script_id.to_string().as_str())
+    );
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn cx_app_id_forwarded_for_attribution() {
+    let http = FakeHttp::responding(200, "text/plain", "ok");
+    let engine = engine_with(http.clone());
+    let app = AppId::new();
+    let src = r#"http::get("https://x/").status"#;
+    let _ = run(engine, src, baseline_request(app, ScriptId::new())).await;
+    assert_eq!(http.recorded.lock().unwrap().last_app, Some(app));
+}

crates/executor-core/tests/sdk_kv.rs

@@ -10,8 +10,9 @@ use std::sync::Arc;
 use async_trait::async_trait;
 use picloud_executor_core::{Engine, ExecRequest, InvocationType, Limits};
 use picloud_shared::{
-    AppId, ExecutionId, KvError, KvListPage, KvService, NoopDeadLetterService, NoopEventEmitter,
-    RequestId, ScriptId, ScriptSandbox, SdkCallCx, Services,
+    AppId, ExecutionId, KvError, KvListPage, KvService, NoopDeadLetterService, NoopDocsService,
+    NoopEventEmitter, NoopHttpService, NoopModuleSource, RequestId, ScriptId, ScriptSandbox,
+    SdkCallCx, Services,
 };
 use serde_json::{json, Value};
 use tokio::sync::Mutex;
@@ -101,8 +102,13 @@ impl KvService for InMemoryKv {
 fn make_engine() -> Arc<Engine> {
     let services = Services::new(
         Arc::new(InMemoryKv::default()),
+        Arc::new(NoopDocsService),
         Arc::new(NoopDeadLetterService),
         Arc::new(NoopEventEmitter),
+        Arc::new(NoopModuleSource),
+        Arc::new(NoopHttpService),
+        Arc::new(picloud_shared::NoopFilesService),
+        Arc::new(picloud_shared::NoopPubsubService),
     );
     Arc::new(Engine::new(Limits::default(), services))
 }

crates/executor-core/tests/sdk_pubsub.rs

@@ -0,0 +1,157 @@
+//! `pubsub::` SDK bridge integration tests — runs a real Rhai engine
+//! against an in-memory `PubsubService` that records the published
+//! `(topic, message)`. Verifies the message JSON encoding the wire
+//! contract requires: Maps, Arrays, strings, numbers, bool, null, and
+//! **Blob → base64**, including nesting.
+
+use std::collections::BTreeMap;
+use std::sync::{Arc, Mutex};
+
+use async_trait::async_trait;
+use picloud_executor_core::{Engine, ExecRequest, InvocationType, Limits};
+use picloud_shared::{
+    AppId, ExecutionId, NoopDeadLetterService, NoopDocsService, NoopEventEmitter, NoopFilesService,
+    NoopHttpService, NoopKvService, NoopModuleSource, PubsubError, PubsubService, RequestId,
+    ScriptId, ScriptSandbox, SdkCallCx, Services,
+};
+use serde_json::{json, Value};
+
+#[derive(Default)]
+struct RecordingPubsub {
+    last: Mutex<Option<(String, Value)>>,
+}
+
+#[async_trait]
+impl PubsubService for RecordingPubsub {
+    async fn publish_durable(
+        &self,
+        _cx: &SdkCallCx,
+        topic: &str,
+        message: Value,
+    ) -> Result<(), PubsubError> {
+        if topic.trim().is_empty() {
+            return Err(PubsubError::EmptyTopic);
+        }
+        *self.last.lock().unwrap() = Some((topic.to_string(), message));
+        Ok(())
+    }
+}
+
+fn make_engine(svc: Arc<RecordingPubsub>) -> Arc<Engine> {
+    let services = Services::new(
+        Arc::new(NoopKvService),
+        Arc::new(NoopDocsService),
+        Arc::new(NoopDeadLetterService),
+        Arc::new(NoopEventEmitter),
+        Arc::new(NoopModuleSource),
+        Arc::new(NoopHttpService),
+        Arc::new(NoopFilesService),
+        svc,
+    );
+    Arc::new(Engine::new(Limits::default(), services))
+}
+
+fn baseline_request(app_id: AppId) -> ExecRequest {
+    let execution_id = ExecutionId::new();
+    ExecRequest {
+        execution_id,
+        request_id: RequestId::new(),
+        script_id: ScriptId::new(),
+        script_name: "pubsub-test".into(),
+        invocation_type: InvocationType::Http,
+        path: "/pubsub-test".into(),
+        headers: BTreeMap::new(),
+        body: Value::Null,
+        params: BTreeMap::new(),
+        query: BTreeMap::new(),
+        rest: String::new(),
+        sandbox_overrides: ScriptSandbox::default(),
+        app_id,
+        principal: None,
+        trigger_depth: 0,
+        root_execution_id: execution_id,
+        is_dead_letter_handler: false,
+        event: None,
+    }
+}
+
+async fn run(engine: Arc<Engine>, src: &str, req: ExecRequest) {
+    let src = src.to_string();
+    tokio::task::spawn_blocking(move || engine.execute(&src, req))
+        .await
+        .expect("spawn_blocking should not panic")
+        .expect("script execution should succeed");
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn publish_map_message() {
+    let svc = Arc::new(RecordingPubsub::default());
+    let engine = make_engine(svc.clone());
+    run(
+        engine,
+        r#"pubsub::publish_durable("user.created", #{ user_id: "abc", n: 7, ok: true });"#,
+        baseline_request(AppId::new()),
+    )
+    .await;
+    let (topic, msg) = svc.last.lock().unwrap().clone().unwrap();
+    assert_eq!(topic, "user.created");
+    assert_eq!(msg, json!({ "user_id": "abc", "n": 7, "ok": true }));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn publish_scalar_and_array_and_null() {
+    let svc = Arc::new(RecordingPubsub::default());
+    let engine = make_engine(svc.clone());
+    run(
+        engine,
+        r#"pubsub::publish_durable("a", [1, "two", false, ()]);"#,
+        baseline_request(AppId::new()),
+    )
+    .await;
+    let (_t, msg) = svc.last.lock().unwrap().clone().unwrap();
+    assert_eq!(msg, json!([1, "two", false, null]));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn publish_number_scalar() {
+    let svc = Arc::new(RecordingPubsub::default());
+    let engine = make_engine(svc.clone());
+    run(
+        engine,
+        r#"pubsub::publish_durable("metric", 42);"#,
+        baseline_request(AppId::new()),
+    )
+    .await;
+    let (_t, msg) = svc.last.lock().unwrap().clone().unwrap();
+    assert_eq!(msg, json!(42));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn publish_blob_encodes_base64_including_nested() {
+    let svc = Arc::new(RecordingPubsub::default());
+    let engine = make_engine(svc.clone());
+    // base64("hello") = "aGVsbG8=" (STANDARD, padded).
+    run(
+        engine,
+        r#"
+            let data = base64::decode("aGVsbG8=");
+            pubsub::publish_durable("blobs", #{ raw: data, list: [data] });
+        "#,
+        baseline_request(AppId::new()),
+    )
+    .await;
+    let (_t, msg) = svc.last.lock().unwrap().clone().unwrap();
+    assert_eq!(msg, json!({ "raw": "aGVsbG8=", "list": ["aGVsbG8="] }));
+}
+
+#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
+async fn publish_empty_topic_throws() {
+    let svc = Arc::new(RecordingPubsub::default());
+    let engine = make_engine(svc.clone());
+    let src = r#"pubsub::publish_durable("", 1);"#.to_string();
+    let req = baseline_request(AppId::new());
+    let res = tokio::task::spawn_blocking(move || engine.execute(&src, req))
+        .await
+        .expect("spawn_blocking should not panic");
+    assert!(res.is_err(), "empty topic should throw");
+}

crates/manager-core/Cargo.toml

@@ -23,8 +23,11 @@ tokio.workspace = true
 tracing.workspace = true
 uuid.workspace = true
 chrono.workspace = true
+chrono-tz.workspace = true
+cron.workspace = true
 sqlx.workspace = true
 url.workspace = true
+reqwest.workspace = true
 
 argon2.workspace = true
 sha2.workspace = true

crates/manager-core/migrations/0013_docs.sql

@@ -0,0 +1,39 @@
+-- v1.1.2: Documents — schemaless JSONB store with basic query semantics.
+--
+-- Identity tuple `(app_id, collection, id)`. `id` is a server-generated
+-- UUID; scripts never supply it on create. `app_id` is first in the
+-- primary key so the implicit index is always per-app — cross-app reads
+-- are impossible even under a buggy query.
+--
+-- `data` is JSONB so scripts can store nested structures without a
+-- separate serialization step. The GIN-on-`jsonb_path_ops` index
+-- accelerates the v1.1.2 query DSL's equality and containment operators
+-- (`docs::find` with `$eq` / `$in`); range/comparison operators rely on
+-- the per-collection seq scan within the small `app_id` partition.
+--
+-- `created_at` / `updated_at` are server-managed: created on insert,
+-- bumped on every successful update. The returned doc envelope surfaces
+-- both fields to scripts for read-only access (no script-side override).
+
+CREATE TABLE docs (
+    app_id      UUID NOT NULL REFERENCES apps(id) ON DELETE CASCADE,
+    collection  TEXT NOT NULL,
+    id          UUID NOT NULL,
+    data        JSONB NOT NULL,
+    created_at  TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    updated_at  TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    PRIMARY KEY (app_id, collection, id)
+);
+
+-- The dispatcher/find hot path: "all docs in app X / collection Y."
+-- The PK already covers (app_id, collection) as a prefix but spelling
+-- out the explicit index makes intent clear for the planner. Mirrors
+-- 0007_kv.sql's idx_kv_entries_app_collection.
+CREATE INDEX idx_docs_app_collection ON docs (app_id, collection);
+
+-- GIN on JSONB with the `jsonb_path_ops` opclass: smaller index than
+-- the default `jsonb_ops`, supports `@>` (containment) which is what
+-- equality filters compile to under the GIN-friendly path. Range
+-- operators ($gt/$gte/$lt/$lte/$ne) fall back to per-collection scans;
+-- those are still bounded by the (app_id, collection) selectivity.
+CREATE INDEX idx_docs_data_gin ON docs USING GIN (data jsonb_path_ops);

crates/manager-core/migrations/0014_docs_triggers.sql

@@ -0,0 +1,36 @@
+-- v1.1.2: Extend the triggers framework to recognise `docs` as the
+-- second concrete kind (after `kv` in v1.1.1).
+--
+-- Two CHECK constraints widen (no narrowing — both lists strictly
+-- gain `'docs'`); one new detail table mirrors `kv_trigger_details`'s
+-- shape with `DocsEventOp` ops instead of `KvEventOp`. Dispatcher
+-- routing is generic across kinds — the same code path that handles
+-- `Kv | DeadLetter` outbox rows now also handles `Docs` (single match
+-- arm extension on the Rust side; no migration needed).
+
+-- Extend triggers.kind to include 'docs'. Constraint is in-line on the
+-- column so Postgres auto-named it `triggers_kind_check`. Dropping the
+-- old and adding the widened constraint is safe — no existing rows
+-- carry a value outside the new set.
+ALTER TABLE triggers DROP CONSTRAINT triggers_kind_check;
+ALTER TABLE triggers ADD CONSTRAINT triggers_kind_check
+    CHECK (kind IN ('kv', 'dead_letter', 'docs'));
+
+-- Extend outbox.source_kind to include 'docs'. Same shape as above;
+-- v1.1.1's existing source_kinds ('http', 'kv', 'dead_letter') stay.
+ALTER TABLE outbox DROP CONSTRAINT outbox_source_kind_check;
+ALTER TABLE outbox ADD CONSTRAINT outbox_source_kind_check
+    CHECK (source_kind IN ('http', 'kv', 'dead_letter', 'docs'));
+
+-- One row per docs trigger. Same shape as `kv_trigger_details`:
+--   collection_glob — "*" matches all, "foo*" prefix-matches, "foo"
+--                     exact-matches (Rust-side via collection_matches).
+--   ops             — subset of {create, update, delete}. Empty array
+--                     means "any op" (matches every docs mutation in
+--                     the collection). The admin endpoint rejects
+--                     empty collection_glob; ops can be empty.
+CREATE TABLE docs_trigger_details (
+    trigger_id        UUID PRIMARY KEY REFERENCES triggers(id) ON DELETE CASCADE,
+    collection_glob   TEXT NOT NULL,
+    ops               TEXT[] NOT NULL
+);

crates/manager-core/migrations/0015_scripts_kind.sql

@@ -0,0 +1,31 @@
+-- v1.1.3: distinguish endpoint scripts (HTTP / trigger entry points) from
+-- module scripts (libraries `import`ed by other scripts). The Rhai module
+-- resolver added in v1.1.3 looks up `kind = 'module'` rows by
+-- `(app_id, name)`; route bind and trigger create reject `kind = 'module'`
+-- targets.
+--
+-- Backfill: existing rows take the DEFAULT clause on column add. Every
+-- script that existed in v1.0 / v1.1.0 / v1.1.1 / v1.1.2 was an endpoint
+-- (the only kind those versions supported), which matches the default.
+ALTER TABLE scripts
+    ADD COLUMN kind TEXT NOT NULL DEFAULT 'endpoint'
+        CHECK (kind IN ('endpoint', 'module'));
+
+-- Composite index on (app_id, kind) so the resolver's per-app module
+-- lookup ("modules in app X named Y") is one index scan. The existing
+-- per-app UNIQUE on `name` already serves name-based lookups, but it
+-- doesn't help when filtering specifically for `kind = 'module'`.
+CREATE INDEX idx_scripts_app_kind ON scripts (app_id, kind);
+
+-- Modules are imported by exact string name; arbitrary spaces / control
+-- characters would make `import "<name>"` fragile. We constrain module
+-- names to a conservative identifier shape (letters, digits, underscore;
+-- starts with a non-digit; up to 64 chars). Endpoint scripts keep the
+-- looser pre-v1.1.3 name rules — the dashboard generates endpoint names
+-- (and some users may already have spaces in them; we don't break those).
+ALTER TABLE scripts
+    ADD CONSTRAINT scripts_module_name_shape
+    CHECK (
+        kind <> 'module'
+        OR name ~ '^[a-zA-Z_][a-zA-Z0-9_]{0,63}$'
+    );

crates/manager-core/migrations/0016_script_imports.sql

@@ -0,0 +1,35 @@
+-- v1.1.3: dep graph between scripts and the modules they `import`.
+--
+-- Populated at script save-time. The validator extracts literal-path
+-- `import "<name>"` declarations from the AST; the script repo writes
+-- one row per resolved (importer, imported) pair inside the same
+-- transaction as the INSERT/UPDATE on `scripts`. Unresolved names
+-- (imported module doesn't exist yet) are silently skipped — the
+-- resolver returns ErrorModuleNotFound at runtime, and a later save
+-- of either script re-resolves and writes the edge.
+--
+-- Dynamic imports (`import some_var as alias;`) are not tracked
+-- here — the resolver still honors them at runtime, but the graph
+-- only captures names known at compile time. Document as a known
+-- v1.1.3 limitation.
+--
+-- Purpose: drives a future "Used by" panel on a module's detail page
+-- (v1.2+) and is the foundation for cluster-mode eager cache
+-- invalidation (v1.3+). v1.1.3 only persists the rows; no admin
+-- endpoint surfaces them yet.
+CREATE TABLE script_imports (
+    app_id              UUID NOT NULL REFERENCES apps(id) ON DELETE CASCADE,
+    importer_script_id  UUID NOT NULL REFERENCES scripts(id) ON DELETE CASCADE,
+    imported_script_id  UUID NOT NULL REFERENCES scripts(id) ON DELETE CASCADE,
+    created_at          TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    PRIMARY KEY (importer_script_id, imported_script_id)
+);
+
+-- Reverse-edge index: "list scripts that import module X". The PK
+-- covers (importer, imported) so forward lookups by importer are
+-- already free; the reverse direction needs its own index.
+CREATE INDEX idx_script_imports_imported ON script_imports (imported_script_id);
+
+-- App-scoped scan ("all imports in this app") — used by the schema
+-- snapshot tests and (eventually) the admin "audit" view.
+CREATE INDEX idx_script_imports_app ON script_imports (app_id);

crates/manager-core/migrations/0017_cron_triggers.sql

@@ -0,0 +1,43 @@
+-- v1.1.4: Extend the triggers framework to recognise `cron` as the
+-- fourth concrete kind (after `kv` v1.1.1, `dead_letter` v1.1.1, `docs`
+-- v1.1.2). Mirrors the 0014 docs extension: two CHECK constraints widen
+-- (strictly gaining `'cron'`), one new detail table.
+--
+-- Cron rows route through the SAME generic dispatcher path as kv/docs/
+-- dead_letter (single match-arm extension on the Rust side). The only
+-- new machinery is a scheduler task that enqueues due cron triggers
+-- into the outbox; dispatch itself is unchanged.
+
+-- Extend triggers.kind to include 'cron'. No existing row carries a
+-- value outside the widened set, so the drop+add is safe.
+ALTER TABLE triggers DROP CONSTRAINT triggers_kind_check;
+ALTER TABLE triggers ADD CONSTRAINT triggers_kind_check
+    CHECK (kind IN ('kv', 'dead_letter', 'docs', 'cron'));
+
+-- Extend outbox.source_kind to include 'cron'. v1.1.x's existing
+-- source_kinds ('http', 'kv', 'dead_letter', 'docs') stay.
+ALTER TABLE outbox DROP CONSTRAINT outbox_source_kind_check;
+ALTER TABLE outbox ADD CONSTRAINT outbox_source_kind_check
+    CHECK (source_kind IN ('http', 'kv', 'dead_letter', 'docs', 'cron'));
+
+-- One row per cron trigger.
+--   schedule      — 6-field cron expression (with seconds), validated
+--                   at insert time by the `cron` crate.
+--   timezone      — IANA tz name (e.g. "America/Los_Angeles"), validated
+--                   via chrono-tz. Required so schedules like "every
+--                   weekday at 9am" are unambiguous. Defaults to UTC.
+--   last_fired_at — set transactionally with each enqueue. NULL until
+--                   the trigger first fires. The scheduler computes the
+--                   next fire time in-process from
+--                   (schedule, timezone, last_fired_at); there is no
+--                   stored next_fire column (kept stateless on purpose).
+CREATE TABLE cron_trigger_details (
+    trigger_id    UUID PRIMARY KEY REFERENCES triggers(id) ON DELETE CASCADE,
+    schedule      TEXT NOT NULL,
+    timezone      TEXT NOT NULL DEFAULT 'UTC',
+    last_fired_at TIMESTAMPTZ
+);
+
+-- Hot lookup for the scheduler: "all enabled cron triggers due now"
+-- scans by last_fired_at.
+CREATE INDEX idx_cron_triggers_due ON cron_trigger_details (last_fired_at);

crates/manager-core/migrations/0018_files.sql

@@ -0,0 +1,25 @@
+-- v1.1.5: filesystem-backed blob storage. The row holds metadata +
+-- the SHA-256 checksum; the blob bytes live on disk at
+--   <PICLOUD_FILES_ROOT>/files/<app_id>/<collection>/<id[0:2]>/<id>
+-- (never in Postgres). Identity tuple is (app_id, collection, id) per
+-- docs/sdk-shape.md, matching KV/docs collection scoping.
+--
+-- The checksum is computed in a single pass during the atomic write and
+-- re-verified on read (FilesError::Corrupted on mismatch). Per-app
+-- quotas are deferred to v1.2; only the per-file size cap is enforced
+-- (in the service, not the schema).
+CREATE TABLE files (
+    app_id          UUID NOT NULL REFERENCES apps(id) ON DELETE CASCADE,
+    collection      TEXT NOT NULL,
+    id              UUID NOT NULL,
+    name            TEXT NOT NULL,
+    content_type    TEXT NOT NULL,
+    size_bytes      BIGINT NOT NULL,
+    checksum_sha256 TEXT NOT NULL,           -- hex, 64 chars, lowercase
+    created_at      TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    updated_at      TIMESTAMPTZ NOT NULL DEFAULT NOW(),
+    PRIMARY KEY (app_id, collection, id)
+);
+
+-- List + cursor pagination scans by (app_id, collection).
+CREATE INDEX idx_files_app_collection ON files (app_id, collection);

crates/manager-core/migrations/0019_files_triggers.sql

@@ -0,0 +1,29 @@
+-- v1.1.5: extend the triggers framework to recognise `files` as the
+-- fifth concrete kind (after `kv`/`dead_letter` v1.1.1, `docs` v1.1.2,
+-- `cron` v1.1.4). Mirrors the 0014/0017 extensions exactly: two CHECK
+-- constraints widen (strictly gaining `'files'`), one new detail table.
+--
+-- Files rows route through the SAME generic dispatcher path as the
+-- other event kinds (single match-arm extension on the Rust side). The
+-- only new machinery is the FilesServiceImpl emitting ServiceEvents
+-- that the OutboxEventEmitter fans out — identical to KV/docs.
+
+-- Extend triggers.kind to include 'files'. No existing row carries a
+-- value outside the widened set, so the drop+add is safe.
+ALTER TABLE triggers DROP CONSTRAINT triggers_kind_check;
+ALTER TABLE triggers ADD CONSTRAINT triggers_kind_check
+    CHECK (kind IN ('kv', 'dead_letter', 'docs', 'cron', 'files'));
+
+-- Extend outbox.source_kind to include 'files'.
+ALTER TABLE outbox DROP CONSTRAINT outbox_source_kind_check;
+ALTER TABLE outbox ADD CONSTRAINT outbox_source_kind_check
+    CHECK (source_kind IN ('http', 'kv', 'dead_letter', 'docs', 'cron', 'files'));
+
+-- One row per files trigger. Mirrors kv_trigger_details:
+--   collection_glob — "*", "exact", or "prefix*"
+--   ops             — subset of {create, update, delete}, empty = any
+CREATE TABLE files_trigger_details (
+    trigger_id        UUID PRIMARY KEY REFERENCES triggers(id) ON DELETE CASCADE,
+    collection_glob   TEXT NOT NULL,
+    ops               TEXT[] NOT NULL
+);

crates/manager-core/migrations/0020_pubsub_triggers.sql

@@ -0,0 +1,34 @@
+-- v1.1.5: extend the triggers framework to recognise `pubsub` as the
+-- sixth concrete kind. Same Layout-E shape as files (0019): two CHECK
+-- constraints widen, one new detail table.
+--
+-- Pub/sub fans out at PUBLISH time (one outbox row per matching trigger,
+-- written by the PubsubServiceImpl), so the dispatcher needs no pubsub-
+-- specific branching — a pubsub outbox row dispatches like any other
+-- async trigger.
+
+-- Extend triggers.kind to include 'pubsub'.
+ALTER TABLE triggers DROP CONSTRAINT triggers_kind_check;
+ALTER TABLE triggers ADD CONSTRAINT triggers_kind_check
+    CHECK (kind IN ('kv', 'dead_letter', 'docs', 'cron', 'files', 'pubsub'));
+
+-- Extend outbox.source_kind to include 'pubsub'.
+ALTER TABLE outbox DROP CONSTRAINT outbox_source_kind_check;
+ALTER TABLE outbox ADD CONSTRAINT outbox_source_kind_check
+    CHECK (source_kind IN ('http', 'kv', 'dead_letter', 'docs',
+                           'cron', 'files', 'pubsub'));
+
+-- One row per pubsub trigger. `topic_pattern` is "exact", "prefix.*",
+-- or "*" — validated in Rust at trigger creation. Topics are implicit
+-- on first publish; the external-subscribable `topics` table is v1.1.6.
+CREATE TABLE pubsub_trigger_details (
+    trigger_id     UUID PRIMARY KEY REFERENCES triggers(id) ON DELETE CASCADE,
+    topic_pattern  TEXT NOT NULL
+);
+
+-- Hot lookup for fan-out: "all enabled pubsub triggers in app X".
+-- Third partial index of its kind (after v1.1.1's idx_triggers_app_kind_
+-- enabled); partial indexes are tiny and the planner picks the narrowest.
+CREATE INDEX idx_triggers_app_pubsub_enabled
+    ON triggers (app_id, kind)
+    WHERE enabled = TRUE AND kind = 'pubsub';

crates/manager-core/src/api.rs

@@ -12,8 +12,8 @@ use axum::{
     Extension, Json, Router,
 };
 use picloud_shared::{
-    AppId, ExecutionLog, InstanceRole, Principal, Script, ScriptId, ScriptSandbox, ScriptValidator,
-    ValidationError,
+    AppId, ExecutionLog, InstanceRole, Principal, Script, ScriptId, ScriptKind, ScriptSandbox,
+    ScriptValidator, ValidatedScript, ValidationError,
 };
 use serde::Deserialize;
 
@@ -88,6 +88,11 @@ pub struct CreateScriptRequest {
     pub name: String,
     pub description: Option<String>,
     pub source: String,
+    /// v1.1.3: `endpoint` (default — handles HTTP routes / trigger
+    /// targets) or `module` (library of fn/const imported by other
+    /// scripts). Modules reject route binding and trigger creation.
+    #[serde(default)]
+    pub kind: ScriptKind,
     pub timeout_seconds: Option<i32>,
     pub memory_limit_mb: Option<i32>,
     /// Sandbox overrides; absent or empty `{}` means "use platform
@@ -120,6 +125,10 @@ pub struct UpdateScriptRequest {
     /// `Some(ScriptSandbox::empty())` to clear them). Absent leaves
     /// the stored value unchanged.
     pub sandbox: Option<ScriptSandbox>,
+    /// v1.1.3: `Some(kind)` changes the script's role. Transitions to
+    /// `Module` are rejected if any routes or triggers still reference
+    /// the script. `module → endpoint` is always allowed.
+    pub kind: Option<ScriptKind>,
 }
 
 #[allow(clippy::option_option)]
@@ -202,7 +211,20 @@ async fn create_script<R: ScriptRepository, L: ExecutionLogRepository>(
         Capability::AppWriteScript(input.app_id),
     )
     .await?;
-    state.validator.validate(&input.source)?;
+    // v1.1.3: dispatch to the right validator based on declared kind.
+    // Module bodies have stricter rules (no top-level statements) so
+    // they need a separate gate; endpoints retain the parse-only path.
+    let validated: ValidatedScript = if input.kind == ScriptKind::Module {
+        if RESERVED_MODULE_NAMES.contains(&input.name.as_str()) {
+            return Err(ApiError::Invalid(ValidationError::ModuleShape(format!(
+                "{:?} is a reserved module name (shadows a built-in SDK namespace)",
+                input.name
+            ))));
+        }
+        state.validator.validate_module(&input.source)?
+    } else {
+        state.validator.validate(&input.source)?
+    };
     state.sandbox_ceiling.check(&input.sandbox)?;
     // Refuse early if the app_id doesn't exist — a clean 422 beats a
     // raw FK violation surfacing as 500.
@@ -216,6 +238,7 @@ async fn create_script<R: ScriptRepository, L: ExecutionLogRepository>(
             name: input.name,
             description: input.description,
             source: input.source,
+            kind: input.kind,
             timeout_seconds: input.timeout_seconds,
             memory_limit_mb: input.memory_limit_mb,
             sandbox: if input.sandbox.is_empty() {
@@ -223,11 +246,39 @@ async fn create_script<R: ScriptRepository, L: ExecutionLogRepository>(
             } else {
                 Some(input.sandbox)
             },
+            imports: validated.imports,
         })
         .await?;
     Ok((StatusCode::CREATED, Json(created)))
 }
 
+/// Module names that would shadow a built-in stdlib / service namespace.
+/// Rejected at create time so `import "kv" as foo` can never resolve to
+/// a user-supplied module instead of (in a hypothetical future) the
+/// real KV bridge — defense against author confusion, not a security
+/// boundary (stdlib namespaces and module imports already live in
+/// disjoint Rhai scopes).
+const RESERVED_MODULE_NAMES: &[&str] = &[
+    "log",
+    "regex",
+    "random",
+    "time",
+    "json",
+    "base64",
+    "hex",
+    "url",
+    "kv",
+    "docs",
+    "dead_letters",
+    "http",
+    "files",
+    "pubsub",
+    "secrets",
+    "email",
+    "users",
+    "queue",
+];
+
 async fn update_script<R: ScriptRepository, L: ExecutionLogRepository>(
     State(state): State<AdminState<R, L>>,
     Extension(principal): Extension<Principal>,
@@ -241,9 +292,44 @@ async fn update_script<R: ScriptRepository, L: ExecutionLogRepository>(
         Capability::AppWriteScript(script.app_id),
     )
     .await?;
-    if let Some(src) = input.source.as_deref() {
-        state.validator.validate(src)?;
+
+    // Effective post-update kind: explicit override > existing kind.
+    let effective_kind = input.kind.unwrap_or(script.kind);
+
+    // v1.1.3: reject `endpoint → module` if the script still has
+    // routes or triggers bound to it. The reverse direction is always
+    // allowed (a module can't have routes/triggers anyway, so the
+    // transition can never strand users).
+    if effective_kind == ScriptKind::Module && script.kind != ScriptKind::Module {
+        let routes = state.repo.count_routes_for_script(id).await?;
+        let triggers = state.repo.count_triggers_for_script(id).await?;
+        if routes + triggers > 0 {
+            return Err(ApiError::Invalid(ValidationError::ModuleShape(format!(
+                "cannot change kind to module: script is referenced by {routes} route(s) and {triggers} trigger(s); detach them first"
+            ))));
         }
+        if RESERVED_MODULE_NAMES.contains(&script.name.as_str()) {
+            return Err(ApiError::Invalid(ValidationError::ModuleShape(format!(
+                "{:?} is a reserved module name (shadows a built-in SDK namespace)",
+                script.name
+            ))));
+        }
+    }
+
+    // v1.1.3: re-validate using the effective kind so endpoint → module
+    // transitions with a fresh source enforce the module shape rules.
+    // Source-less edits (name/description only) don't re-validate.
+    let imports_for_patch: Option<Vec<String>> = if let Some(src) = input.source.as_deref() {
+        let validated = if effective_kind == ScriptKind::Module {
+            state.validator.validate_module(src)?
+        } else {
+            state.validator.validate(src)?
+        };
+        Some(validated.imports)
+    } else {
+        None
+    };
+
     if let Some(sb) = input.sandbox.as_ref() {
         state.sandbox_ceiling.check(sb)?;
     }
@@ -258,6 +344,8 @@ async fn update_script<R: ScriptRepository, L: ExecutionLogRepository>(
                 timeout_seconds: input.timeout_seconds,
                 memory_limit_mb: input.memory_limit_mb,
                 sandbox: input.sandbox,
+                kind: input.kind,
+                imports: imports_for_patch,
             },
         )
         .await?;

crates/manager-core/src/app_bootstrap.rs

@@ -64,9 +64,11 @@ async fn seed_into(
             name: "hello".to_string(),
             description: Some("Reference example: returns a greeting at GET /hello.".to_string()),
             source: HELLO_RHAI_SOURCE.to_string(),
+            kind: picloud_shared::ScriptKind::Endpoint,
             timeout_seconds: Some(5),
             memory_limit_mb: None,
             sandbox: None,
+            imports: Vec::new(),
         })
         .await?;
 

crates/manager-core/src/authz.rs

@@ -64,6 +64,31 @@ pub enum Capability {
     /// Write entries to this app's KV store (v1.1.1). Granted to
     /// `editor`+. Maps to `script:write` on API keys.
     AppKvWrite(AppId),
+    /// Read documents from this app's docs store (v1.1.2). Same trust
+    /// shape as KV read — granted to `viewer`+, maps to `script:read`
+    /// on API keys. Honors the seven-scope commitment.
+    AppDocsRead(AppId),
+    /// Write documents to this app's docs store (v1.1.2). Same trust
+    /// shape as KV write — granted to `editor`+, maps to
+    /// `script:write` on API keys.
+    AppDocsWrite(AppId),
+    /// Make an outbound HTTP request from a script in this app
+    /// (v1.1.4). Maps to `script:write` on API keys: any outbound
+    /// request can exfiltrate data — including read methods like GET —
+    /// so the conservative write mapping is correct. Splitting
+    /// read/write is a v1.2+ refinement. Granted to `editor`+.
+    AppHttpRequest(AppId),
+    /// Read blobs from this app's files store (v1.1.5). Same trust
+    /// shape as KV/docs read — granted to `viewer`+, maps to
+    /// `script:read` on API keys. Honors the seven-scope commitment.
+    AppFilesRead(AppId),
+    /// Write blobs to this app's files store (v1.1.5). Granted to
+    /// `editor`+, maps to `script:write` on API keys.
+    AppFilesWrite(AppId),
+    /// Publish a durable pub/sub message from a script in this app
+    /// (v1.1.5). Maps to `script:write` on API keys (a publish is a
+    /// write that fans out to subscribers). Granted to `editor`+.
+    AppPubsubPublish(AppId),
     /// Create / list / delete triggers for this app (v1.1.1). Maps to
     /// `app:admin` on API keys — triggers are app-configuration acts
     /// rather than data-plane access. Granted to `app_admin`+.
@@ -91,6 +116,12 @@ impl Capability {
             | Self::AppLogRead(id)
             | Self::AppKvRead(id)
             | Self::AppKvWrite(id)
+            | Self::AppDocsRead(id)
+            | Self::AppDocsWrite(id)
+            | Self::AppHttpRequest(id)
+            | Self::AppFilesRead(id)
+            | Self::AppFilesWrite(id)
+            | Self::AppPubsubPublish(id)
             | Self::AppManageTriggers(id)
             | Self::AppDeadLetterManage(id) => Some(id),
         }
@@ -107,8 +138,16 @@ impl Capability {
             Self::InstanceCreateApp | Self::InstanceManageUsers | Self::InstanceManageSettings => {
                 Scope::InstanceAdmin
             }
-            Self::AppRead(_) | Self::AppKvRead(_) => Scope::ScriptRead,
-            Self::AppWriteScript(_) | Self::AppKvWrite(_) => Scope::ScriptWrite,
+            Self::AppRead(_)
+            | Self::AppKvRead(_)
+            | Self::AppDocsRead(_)
+            | Self::AppFilesRead(_) => Scope::ScriptRead,
+            Self::AppWriteScript(_)
+            | Self::AppKvWrite(_)
+            | Self::AppDocsWrite(_)
+            | Self::AppHttpRequest(_)
+            | Self::AppFilesWrite(_)
+            | Self::AppPubsubPublish(_) => Scope::ScriptWrite,
             Self::AppWriteRoute(_) => Scope::RouteWrite,
             Self::AppManageDomains(_) => Scope::DomainManage,
             Self::AppAdmin(_) | Self::AppManageTriggers(_) | Self::AppDeadLetterManage(_) => {
@@ -253,7 +292,11 @@ async fn member_grants(
 const fn role_satisfies(role: AppRole, cap: Capability) -> bool {
     let in_viewer = matches!(
         cap,
-        Capability::AppRead(_) | Capability::AppLogRead(_) | Capability::AppKvRead(_)
+        Capability::AppRead(_)
+            | Capability::AppLogRead(_)
+            | Capability::AppKvRead(_)
+            | Capability::AppDocsRead(_)
+            | Capability::AppFilesRead(_)
     );
     let in_editor = in_viewer
         || matches!(
@@ -261,6 +304,10 @@ const fn role_satisfies(role: AppRole, cap: Capability) -> bool {
             Capability::AppWriteScript(_)
                 | Capability::AppWriteRoute(_)
                 | Capability::AppKvWrite(_)
+                | Capability::AppDocsWrite(_)
+                | Capability::AppHttpRequest(_)
+                | Capability::AppFilesWrite(_)
+                | Capability::AppPubsubPublish(_)
         );
     let in_app_admin = in_editor
         || matches!(

crates/manager-core/src/cron_scheduler.rs

@@ -0,0 +1,297 @@
+//! Cron scheduler — the v1.1.4 time-based trigger source.
+//!
+//! A single tokio task polls `cron_trigger_details` on a tick (default
+//! 30s; `PICLOUD_CRON_TICK_INTERVAL_MS`). For each enabled cron trigger
+//! whose next scheduled fire is due, it enqueues ONE outbox row
+//! (`source_kind = 'cron'`) and updates `last_fired_at` — both in the
+//! same transaction, claimed via `FOR UPDATE SKIP LOCKED` so a future
+//! multi-node deploy can't double-fire.
+//!
+//! The scheduler does NOT dispatch or touch the `ExecutionGate`: it only
+//! enqueues. The existing dispatcher picks the row up and acquires the
+//! gate exactly as it does for kv/docs/dead_letter rows.
+//!
+//! **Catch-up policy (matches the brief):** a trigger that missed N fire
+//! windows since `last_fired_at` fires exactly ONCE on the next tick,
+//! not N times. This falls out of the design: [`next_due`] returns a
+//! single canonical scheduled time (the first slot after the reference
+//! point), and after firing we set `last_fired_at = now`, so the next
+//! tick computes from `now` and sees only future slots. Backfilling
+//! missed windows is intentionally out of scope (an explicit replay
+//! action is the v1.2+ escape hatch).
+
+use std::str::FromStr;
+use std::time::Duration;
+
+use chrono::{DateTime, Utc};
+use chrono_tz::Tz;
+use cron::Schedule;
+use picloud_shared::TriggerEvent;
+use sqlx::PgPool;
+use uuid::Uuid;
+
+/// Validate a 6-field cron expression. Returns the parse error message
+/// on failure.
+///
+/// # Errors
+///
+/// Returns the underlying parse error string when `schedule` is not a
+/// valid cron expression.
+pub fn validate_schedule(schedule: &str) -> Result<(), String> {
+    Schedule::from_str(schedule)
+        .map(|_| ())
+        .map_err(|e| e.to_string())
+}
+
+/// Validate an IANA timezone name (e.g. `America/Los_Angeles`).
+///
+/// # Errors
+///
+/// Returns an error string when `timezone` is not a known IANA name.
+pub fn validate_timezone(timezone: &str) -> Result<(), String> {
+    Tz::from_str(timezone)
+        .map(|_| ())
+        .map_err(|_| format!("unknown IANA timezone: {timezone}"))
+}
+
+/// Compute whether a cron trigger is due, and if so its canonical
+/// scheduled-at moment (UTC).
+///
+/// Returns `Some(scheduled_at)` when the first scheduled slot after the
+/// reference point (`last_fired_at`, or `created_at` if never fired) is
+/// at/before `now`; `None` otherwise. Returns `None` if the schedule or
+/// timezone fails to parse (the row is skipped — it should never have
+/// been inserted, since the admin endpoint validates).
+#[must_use]
+pub fn next_due(
+    schedule: &str,
+    timezone: &str,
+    last_fired_at: Option<DateTime<Utc>>,
+    created_at: DateTime<Utc>,
+    now: DateTime<Utc>,
+) -> Option<DateTime<Utc>> {
+    let sched = Schedule::from_str(schedule).ok()?;
+    let tz = Tz::from_str(timezone).ok()?;
+    // Reference: the last actual fire, or creation if never fired. A
+    // never-fired trigger fires at its first slot at/after creation.
+    let base = last_fired_at.unwrap_or(created_at);
+    let base_tz = base.with_timezone(&tz);
+    let next = sched.after(&base_tz).next()?;
+    let next_utc = next.with_timezone(&Utc);
+    (next_utc <= now).then_some(next_utc)
+}
+
+/// Spawn the scheduler loop. Runs for the process lifetime.
+pub fn spawn_cron_scheduler(pool: PgPool, tick_interval_ms: u32) {
+    // Floor the tick at 1s so a misconfigured 0 can't spin.
+    let interval = Duration::from_millis(u64::from(tick_interval_ms).max(1_000));
+    tokio::spawn(async move {
+        let mut ticker = tokio::time::interval(interval);
+        // Skip the immediate first fire so we don't race startup.
+        ticker.tick().await;
+        loop {
+            ticker.tick().await;
+            if let Err(e) = tick(&pool, Utc::now()).await {
+                tracing::warn!(?e, "cron scheduler tick errored");
+            }
+        }
+    });
+}
+
+#[derive(sqlx::FromRow)]
+struct DueRow {
+    id: Uuid,
+    app_id: Uuid,
+    script_id: Uuid,
+    registered_by_principal: Uuid,
+    created_at: DateTime<Utc>,
+    schedule: String,
+    timezone: String,
+    last_fired_at: Option<DateTime<Utc>>,
+}
+
+/// One scheduler tick: claim enabled cron rows, enqueue the due ones,
+/// bump `last_fired_at`. Returns the number of triggers fired.
+async fn tick(pool: &PgPool, now: DateTime<Utc>) -> Result<usize, sqlx::Error> {
+    let mut tx = pool.begin().await?;
+    let rows: Vec<DueRow> = sqlx::query_as(
+        "SELECT t.id, t.app_id, t.script_id, t.registered_by_principal, t.created_at, \
+                d.schedule, d.timezone, d.last_fired_at \
+         FROM cron_trigger_details d \
+         JOIN triggers t ON t.id = d.trigger_id \
+         WHERE t.enabled = TRUE \
+         FOR UPDATE OF d SKIP LOCKED",
+    )
+    .fetch_all(&mut *tx)
+    .await?;
+
+    let mut fired = 0usize;
+    for r in rows {
+        let Some(scheduled_at) =
+            next_due(&r.schedule, &r.timezone, r.last_fired_at, r.created_at, now)
+        else {
+            continue;
+        };
+
+        let event = TriggerEvent::Cron {
+            schedule: r.schedule.clone(),
+            timezone: r.timezone.clone(),
+            scheduled_at,
+            fired_at: now,
+        };
+        let payload = serde_json::to_value(&event)
+            .map_err(|e| sqlx::Error::Decode(Box::new(std::io::Error::other(e))))?;
+
+        // Enqueue exactly one outbox row. Relies on the same column
+        // defaults the OutboxEventEmitter uses (next_attempt_at = NOW(),
+        // attempt_count = 0, claimed_at NULL → immediately due).
+        sqlx::query(
+            "INSERT INTO outbox \
+                (app_id, source_kind, trigger_id, script_id, payload, \
+                 origin_principal, trigger_depth) \
+             VALUES ($1, 'cron', $2, $3, $4, $5, 0)",
+        )
+        .bind(r.app_id)
+        .bind(r.id)
+        .bind(r.script_id)
+        .bind(payload)
+        .bind(r.registered_by_principal)
+        .execute(&mut *tx)
+        .await?;
+
+        sqlx::query("UPDATE cron_trigger_details SET last_fired_at = $2 WHERE trigger_id = $1")
+            .bind(r.id)
+            .bind(now)
+            .execute(&mut *tx)
+            .await?;
+
+        fired += 1;
+    }
+
+    tx.commit().await?;
+    Ok(fired)
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use chrono::TimeZone;
+
+    #[test]
+    fn valid_six_field_schedule_accepted() {
+        // sec min hour dom mon dow — "every weekday at 9am".
+        validate_schedule("0 0 9 * * MON-FRI").unwrap();
+        validate_schedule("*/5 * * * * *").unwrap();
+        validate_schedule("0 0 0 1 1 *").unwrap();
+    }
+
+    #[test]
+    fn invalid_schedules_rejected() {
+        // 5-field (no seconds) is not the format we accept.
+        assert!(validate_schedule("* * * * *").is_err());
+        // Gibberish.
+        assert!(validate_schedule("not a cron").is_err());
+        assert!(validate_schedule("").is_err());
+        // Out-of-range hour.
+        assert!(validate_schedule("0 0 99 * * *").is_err());
+    }
+
+    #[test]
+    fn known_timezones_accepted() {
+        validate_timezone("UTC").unwrap();
+        validate_timezone("America/Los_Angeles").unwrap();
+        validate_timezone("Europe/Berlin").unwrap();
+    }
+
+    #[test]
+    fn unknown_timezones_rejected() {
+        assert!(validate_timezone("Mars/Phobos").is_err());
+        assert!(validate_timezone("PST").is_err()); // abbreviations aren't IANA names
+        assert!(validate_timezone("").is_err());
+    }
+
+    fn ts(s: &str) -> DateTime<Utc> {
+        DateTime::parse_from_rfc3339(s).unwrap().with_timezone(&Utc)
+    }
+
+    #[test]
+    fn due_when_next_slot_is_at_or_before_now() {
+        // Every minute at second 0. Last fired 90s ago → the next slot
+        // after that is due now.
+        let created = ts("2026-06-01T00:00:00Z");
+        let last = Some(ts("2026-06-15T11:58:10Z"));
+        let now = ts("2026-06-15T12:00:05Z");
+        let due = next_due("0 * * * * *", "UTC", last, created, now);
+        assert_eq!(due, Some(ts("2026-06-15T11:59:00Z")));
+    }
+
+    #[test]
+    fn not_due_when_next_slot_is_in_the_future() {
+        let created = ts("2026-06-01T00:00:00Z");
+        let last = Some(ts("2026-06-15T12:00:00Z"));
+        let now = ts("2026-06-15T12:00:30Z");
+        // Next minute slot is 12:01:00 — still in the future.
+        assert_eq!(next_due("0 * * * * *", "UTC", last, created, now), None);
+    }
+
+    #[test]
+    fn never_fired_uses_created_at_as_reference() {
+        let created = ts("2026-06-15T12:00:10Z");
+        let now = ts("2026-06-15T12:01:30Z");
+        // First slot after creation is 12:01:00, which is <= now → due.
+        let due = next_due("0 * * * * *", "UTC", None, created, now);
+        assert_eq!(due, Some(ts("2026-06-15T12:01:00Z")));
+    }
+
+    /// Catch-up policy: a trigger that missed many windows fires exactly
+    /// ONCE. We simulate two consecutive scheduler ticks the way the DB
+    /// loop does — fire once, set last_fired = now, then re-evaluate.
+    #[test]
+    fn catch_up_fires_exactly_once_after_missed_windows() {
+        let created = ts("2026-06-15T09:00:00Z");
+        // Last fired over 5 minutes (5 windows) ago.
+        let mut last_fired = Some(ts("2026-06-15T11:54:30Z"));
+        let now = ts("2026-06-15T12:00:05Z");
+
+        // Tick 1: due → fire once, advance last_fired to `now`.
+        let first = next_due("0 * * * * *", "UTC", last_fired, created, now);
+        assert!(first.is_some(), "should be due after missing windows");
+        last_fired = Some(now);
+
+        // Tick 2 (same wall-clock): NOT due again — only one fire total,
+        // not one-per-missed-window.
+        let second = next_due("0 * * * * *", "UTC", last_fired, created, now);
+        assert_eq!(second, None, "catch-up must fire exactly once");
+    }
+
+    #[test]
+    fn timezone_affects_fire_time() {
+        // "9am every day" in Los Angeles. On 2026-06-15, PDT = UTC-7, so
+        // 09:00 local = 16:00 UTC.
+        let created = ts("2026-06-15T00:00:00Z");
+        let last = Some(ts("2026-06-15T15:59:00Z"));
+        let now = ts("2026-06-15T16:00:30Z");
+        let due = next_due("0 0 9 * * *", "America/Los_Angeles", last, created, now);
+        assert_eq!(due, Some(ts("2026-06-15T16:00:00Z")));
+        // Sanity: the same expression in UTC would NOT be due at 16:00.
+        assert_eq!(next_due("0 0 9 * * *", "UTC", last, created, now), None);
+    }
+
+    #[test]
+    fn bad_schedule_or_tz_yields_none() {
+        let created = ts("2026-06-15T00:00:00Z");
+        let now = ts("2026-06-15T12:00:00Z");
+        assert_eq!(next_due("garbage", "UTC", None, created, now), None);
+        assert_eq!(
+            next_due("0 * * * * *", "Mars/Phobos", None, created, now),
+            None
+        );
+    }
+
+    #[test]
+    fn utc_offset_constructor_smoke() {
+        // Guard the chrono TimeZone import is actually exercised.
+        let dt = Utc.with_ymd_and_hms(2026, 6, 15, 12, 0, 0).unwrap();
+        assert_eq!(dt, ts("2026-06-15T12:00:00Z"));
+    }
+}

crates/manager-core/src/dead_letters_api.rs

@@ -208,6 +208,9 @@ async fn resolve(
 fn admin_cx(app_id: AppId, principal: &Principal) -> SdkCallCx {
     SdkCallCx {
         app_id,
+        // Admin-plane cx (dead-letter replay/resolve) — no script is
+        // executing, so this attribution id is a fresh sentinel.
+        script_id: picloud_shared::ScriptId::new(),
         principal: Some(principal.clone()),
         execution_id: picloud_shared::ExecutionId::new(),
         request_id: picloud_shared::RequestId::new(),

crates/manager-core/src/dispatcher.rs

@@ -163,7 +163,12 @@ impl Dispatcher {
                     return Ok(());
                 }
             },
-            OutboxSourceKind::Kv | OutboxSourceKind::DeadLetter => {
+            OutboxSourceKind::Kv
+            | OutboxSourceKind::Docs
+            | OutboxSourceKind::DeadLetter
+            | OutboxSourceKind::Cron
+            | OutboxSourceKind::Files
+            | OutboxSourceKind::Pubsub => {
                 let resolved = self.resolve_trigger(&row).await?;
                 let req = match self.build_exec_request(&row, &resolved).await {
                     Ok(req) => req,
@@ -186,9 +191,13 @@ impl Dispatcher {
         // wait synchronously here — sync HTTP and dispatcher share the
         // semaphore so this is intentional.
         let source = resolved.script_source.clone();
+        let identity = picloud_orchestrator_core::ScriptIdentity {
+            script_id: resolved.script_id,
+            updated_at: resolved.script_updated_at,
+        };
         let outcome = self
             .executor
-            .execute(&source, exec_req, ASYNC_EXEC_TIMEOUT)
+            .execute_with_identity(identity, &source, exec_req, ASYNC_EXEC_TIMEOUT)
             .await;
         drop(permit);
 
@@ -230,6 +239,7 @@ impl Dispatcher {
             script_id: script.id,
             script_source: script.source,
             script_name: script.name,
+            script_updated_at: script.updated_at,
             sandbox_overrides: script.sandbox,
             registered_by_principal: trigger.registered_by_principal,
             retry_max_attempts: trigger.retry_max_attempts,
@@ -335,6 +345,7 @@ impl Dispatcher {
             script_id,
             script_source: script.source,
             script_name: payload.script_name,
+            script_updated_at: script.updated_at,
             sandbox_overrides: script.sandbox,
             // HTTP outbox rows don't carry a registered_by_principal
             // — use a sentinel zero UUID since this field isn't used
@@ -516,6 +527,11 @@ pub struct ResolvedTrigger {
     pub script_id: ScriptId,
     pub script_source: String,
     pub script_name: String,
+    /// v1.1.3: freshness comparator for the orchestrator's top-level
+    /// script cache. The dispatcher hands `(script_id, updated_at)`
+    /// in alongside the source so cached ASTs can be reused across
+    /// triggered invocations.
+    pub script_updated_at: chrono::DateTime<chrono::Utc>,
     pub sandbox_overrides: ScriptSandbox,
     pub registered_by_principal: picloud_shared::AdminUserId,
     pub retry_max_attempts: u32,

crates/manager-core/src/docs_filter.rs

@@ -0,0 +1,598 @@
+//! v1.1.2 query DSL parser + AST for `docs::find` / `docs::find_one`.
+//!
+//! Sets the precedent v1.2's `dead_letters::list` will follow (see
+//! `docs/v1.1.x-design-notes.md` §4 #13). When that lands we promote
+//! this module to `picloud-shared` and rename to
+//! `picloud_shared::query::{Filter, FieldPath, ComparisonOp}`; until
+//! then keeping it private to manager-core avoids over-engineering.
+//!
+//! Parse stage is deliberately strict: any unrecognized `$xxx`
+//! operator surfaces as `FilterParseError::UnsupportedOperator` with
+//! a script-visible message naming the offending key + pointing at
+//! v1.2. The error strings become part of the SDK contract once
+//! scripts depend on them; pin them with snapshot tests in the test
+//! module below before changing.
+//!
+//! ## DSL surface (v1.1.2 subset)
+//!
+//! ```rhai
+//! // implicit equality (top-level)
+//! users.find(#{ tier: "gold", status: "active" })
+//!
+//! // operator object on a field
+//! users.find(#{ created_at: #{ "$gt": "2026-01-01T00:00:00Z" } })
+//!
+//! // dotted paths (max 5 segments)
+//! users.find(#{ "user.email": "a@b" })
+//!
+//! // sort + limit as filter modifiers
+//! users.find(#{ tier: "gold", "$sort": #{ created_at: -1 }, "$limit": 10 })
+//! ```
+//!
+//! ## Out of scope (v1.2)
+//!
+//! `$or`, `$and`, `$not`, `$exists`, `$regex`, `$type`, `$size`,
+//! `$all`, `$elemMatch`, multi-field sort, projection, aggregations.
+
+use serde_json::Value;
+
+/// Maximum nesting depth for dotted field paths. `"a.b.c.d.e"` is the
+/// deepest path allowed (5 segments). Deeper paths reject at parse
+/// time with `InvalidFilter` — prevents pathological JSONB navigation
+/// chains from a script.
+pub const MAX_FIELD_PATH_DEPTH: usize = 5;
+
+/// Hard cap on `$limit` values — script-side limits are silently
+/// clamped here so the Postgres query is always bounded. Mirrors the
+/// `find` repo's own internal cap.
+pub const MAX_FIND_LIMIT: u32 = 1_000;
+
+/// Parsed `docs::find` filter.
+#[derive(Debug, Clone, PartialEq)]
+pub struct DocsFilter {
+    pub conditions: Vec<FieldCondition>,
+    pub sort: Option<Sort>,
+    pub limit: Option<u32>,
+}
+
+impl DocsFilter {
+    /// Empty filter — matches every document in the collection.
+    #[must_use]
+    pub const fn empty() -> Self {
+        Self {
+            conditions: Vec::new(),
+            sort: None,
+            limit: None,
+        }
+    }
+}
+
+#[derive(Debug, Clone, PartialEq)]
+pub struct FieldCondition {
+    pub path: FieldPath,
+    pub op: ComparisonOp,
+    pub value: Value,
+}
+
+/// Validated dotted path. Construct only via `FieldPath::parse` so the
+/// segment invariants (non-empty, no `..`, no `$` prefix, depth ≤ 5)
+/// are guaranteed.
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct FieldPath {
+    segments: Vec<String>,
+}
+
+impl FieldPath {
+    /// Parse a dotted path from a JSON object key.
+    pub fn parse(raw: &str) -> Result<Self, FilterParseError> {
+        if raw.is_empty() {
+            return Err(FilterParseError::InvalidFilter(
+                "docs::find: field path must not be empty".into(),
+            ));
+        }
+        let segments: Vec<&str> = raw.split('.').collect();
+        if segments.len() > MAX_FIELD_PATH_DEPTH {
+            return Err(FilterParseError::InvalidFilter(format!(
+                "docs::find: field path '{raw}' exceeds max depth {MAX_FIELD_PATH_DEPTH}"
+            )));
+        }
+        for seg in &segments {
+            if seg.is_empty() {
+                return Err(FilterParseError::InvalidFilter(format!(
+                    "docs::find: field path '{raw}' has an empty segment (leading/trailing dot or '..')"
+                )));
+            }
+            if seg.starts_with('$') {
+                return Err(FilterParseError::InvalidFilter(format!(
+                    "docs::find: field path segment '{seg}' must not start with '$'"
+                )));
+            }
+        }
+        Ok(Self {
+            segments: segments.into_iter().map(ToString::to_string).collect(),
+        })
+    }
+
+    /// Path segments in order. The Postgres impl binds each as a
+    /// separate text parameter to `jsonb_extract_path_text`, so no
+    /// segment ever appears in the SQL string verbatim.
+    #[must_use]
+    pub fn segments(&self) -> &[String] {
+        &self.segments
+    }
+
+    /// Display form for error messages — joined back with `.`.
+    #[must_use]
+    pub fn as_str(&self) -> String {
+        self.segments.join(".")
+    }
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum ComparisonOp {
+    /// Implicit equality at top level OR explicit `$eq`. Maps to
+    /// `jsonb_extract_path_text(...) = $M`.
+    Eq,
+    /// `$ne` — uses Postgres `IS DISTINCT FROM` so JSON nulls and
+    /// missing paths are correctly included (`<>` returns NULL on
+    /// either operand being NULL, which would silently exclude rows
+    /// the user expects to see).
+    Ne,
+    /// `$gt` / `$gte` / `$lt` / `$lte` — text-lex comparison per the
+    /// brief's contract. Known limitation: lex breaks across
+    /// digit-count boundaries (`'10' < '9'` is TRUE). Documented in
+    /// CHANGELOG; v1.2 advanced query will add numeric-aware
+    /// operators.
+    Gt,
+    Gte,
+    Lt,
+    Lte,
+    /// `$in` — `= ANY($M::text[])` where the value list is bound as
+    /// a TEXT[].
+    In,
+}
+
+impl ComparisonOp {
+    /// Decode an operator key like `"$gt"`. Returns `None` for any
+    /// non-`$` key; returns `Some(Err(...))` for `$`-prefixed keys
+    /// not in the v1.1.2 allowlist (caller surfaces the
+    /// UnsupportedOperator error).
+    fn from_dollar_key(key: &str) -> Option<Result<Self, FilterParseError>> {
+        if !key.starts_with('$') {
+            return None;
+        }
+        Some(match key {
+            "$eq" => Ok(Self::Eq),
+            "$ne" => Ok(Self::Ne),
+            "$gt" => Ok(Self::Gt),
+            "$gte" => Ok(Self::Gte),
+            "$lt" => Ok(Self::Lt),
+            "$lte" => Ok(Self::Lte),
+            "$in" => Ok(Self::In),
+            other => Err(FilterParseError::UnsupportedOperator(format!(
+                "docs::find: operator '{other}' is not supported in v1.1.2; planned for v1.2 advanced query"
+            ))),
+        })
+    }
+}
+
+#[derive(Debug, Clone, PartialEq)]
+pub struct Sort {
+    pub path: FieldPath,
+    pub direction: SortDir,
+}
+
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum SortDir {
+    Asc,
+    Desc,
+}
+
+#[derive(Debug, thiserror::Error)]
+pub enum FilterParseError {
+    /// Bad path syntax, malformed operator value, multi-field sort,
+    /// etc. The string is the script-visible message.
+    #[error("{0}")]
+    InvalidFilter(String),
+
+    /// Filter used an operator not in the v1.1.2 allowlist. The
+    /// string includes the offending operator + v1.2 pointer.
+    #[error("{0}")]
+    UnsupportedOperator(String),
+}
+
+/// Parse a `serde_json::Value` filter into `DocsFilter`. The bridge
+/// converts the script's Rhai map into a `Value` via
+/// `executor-core::sdk::bridge::dynamic_to_json` and passes it through
+/// `DocsService::find`; the service calls this parser before touching
+/// the repo.
+pub fn parse_filter(filter: &Value) -> Result<DocsFilter, FilterParseError> {
+    let obj = filter.as_object().ok_or_else(|| {
+        FilterParseError::InvalidFilter("docs::find: filter must be a map/object".into())
+    })?;
+
+    let mut out = DocsFilter::empty();
+
+    for (key, value) in obj {
+        if let Some(stripped) = key.strip_prefix('$') {
+            // Top-level modifier — `$sort` / `$limit`. Any other
+            // dollar-key at top level is unsupported.
+            match stripped {
+                "sort" => out.sort = Some(parse_sort(value)?),
+                "limit" => out.limit = Some(parse_limit(value)?),
+                other => {
+                    return Err(FilterParseError::UnsupportedOperator(format!(
+                        "docs::find: top-level modifier '${other}' is not supported in v1.1.2; planned for v1.2 advanced query"
+                    )));
+                }
+            }
+            continue;
+        }
+
+        // Field path → either implicit equality OR operator-object.
+        let path = FieldPath::parse(key)?;
+        match value {
+            Value::Object(inner) if is_operator_object(inner) => {
+                for (op_key, op_val) in inner {
+                    let Some(op_res) = ComparisonOp::from_dollar_key(op_key) else {
+                        // This shouldn't trigger — is_operator_object
+                        // already guarantees every key is $-prefixed.
+                        return Err(FilterParseError::InvalidFilter(format!(
+                            "docs::find: operator object for '{}' has non-$ key '{op_key}'",
+                            path.as_str()
+                        )));
+                    };
+                    let op = op_res?;
+                    validate_op_value(op, op_val, &path)?;
+                    out.conditions.push(FieldCondition {
+                        path: path.clone(),
+                        op,
+                        value: op_val.clone(),
+                    });
+                }
+            }
+            // Any non-object value is implicit equality.
+            // (Object values with non-$ keys are user data, not an
+            // operator object — reject so the user doesn't accidentally
+            // match against a literal `{ name: "Alice" }` shape that
+            // would never compare meaningfully under JSONB text.)
+            Value::Object(_) => {
+                return Err(FilterParseError::InvalidFilter(format!(
+                    "docs::find: value for '{}' must be a scalar (implicit equality) or an operator map (keys starting with '$')",
+                    path.as_str()
+                )));
+            }
+            _ => {
+                out.conditions.push(FieldCondition {
+                    path,
+                    op: ComparisonOp::Eq,
+                    value: value.clone(),
+                });
+            }
+        }
+    }
+
+    Ok(out)
+}
+
+/// True when every key in the map starts with `$`. Mixed-shape maps
+/// (some `$key`, some user-data key) are rejected to avoid silent
+/// surprise — the user almost certainly meant an operator object.
+fn is_operator_object(map: &serde_json::Map<String, Value>) -> bool {
+    !map.is_empty() && map.keys().all(|k| k.starts_with('$'))
+}
+
+fn validate_op_value(
+    op: ComparisonOp,
+    value: &Value,
+    path: &FieldPath,
+) -> Result<(), FilterParseError> {
+    match op {
+        ComparisonOp::In => {
+            if !value.is_array() {
+                return Err(FilterParseError::InvalidFilter(format!(
+                    "docs::find: '$in' on '{}' requires an array value",
+                    path.as_str()
+                )));
+            }
+        }
+        _ => {
+            // For the scalar-comparison ops, the value must be a JSON
+            // scalar (no arrays / no nested objects). JSON null is
+            // allowed — `$ne` against null is a valid query.
+            if value.is_array() || value.is_object() {
+                return Err(FilterParseError::InvalidFilter(format!(
+                    "docs::find: '{op_name}' on '{path}' requires a scalar value",
+                    op_name = op_name(op),
+                    path = path.as_str()
+                )));
+            }
+        }
+    }
+    Ok(())
+}
+
+const fn op_name(op: ComparisonOp) -> &'static str {
+    match op {
+        ComparisonOp::Eq => "$eq",
+        ComparisonOp::Ne => "$ne",
+        ComparisonOp::Gt => "$gt",
+        ComparisonOp::Gte => "$gte",
+        ComparisonOp::Lt => "$lt",
+        ComparisonOp::Lte => "$lte",
+        ComparisonOp::In => "$in",
+    }
+}
+
+fn parse_sort(value: &Value) -> Result<Sort, FilterParseError> {
+    let map = value.as_object().ok_or_else(|| {
+        FilterParseError::InvalidFilter("docs::find: '$sort' must be a map".into())
+    })?;
+    if map.is_empty() {
+        return Err(FilterParseError::InvalidFilter(
+            "docs::find: '$sort' must name at least one field".into(),
+        ));
+    }
+    if map.len() > 1 {
+        return Err(FilterParseError::InvalidFilter(
+            "docs::find: multi-field '$sort' is not supported in v1.1.2; planned for v1.2 advanced query"
+                .into(),
+        ));
+    }
+    let (field, dir_val) = map.iter().next().unwrap();
+    let path = FieldPath::parse(field)?;
+    let direction = match dir_val.as_i64() {
+        Some(1) => SortDir::Asc,
+        Some(-1) => SortDir::Desc,
+        _ => {
+            return Err(FilterParseError::InvalidFilter(format!(
+                "docs::find: '$sort' direction for '{field}' must be 1 (ascending) or -1 (descending)"
+            )));
+        }
+    };
+    Ok(Sort { path, direction })
+}
+
+fn parse_limit(value: &Value) -> Result<u32, FilterParseError> {
+    let n = value.as_i64().ok_or_else(|| {
+        FilterParseError::InvalidFilter("docs::find: '$limit' must be an integer".into())
+    })?;
+    if n < 0 {
+        return Err(FilterParseError::InvalidFilter(
+            "docs::find: '$limit' must be non-negative".into(),
+        ));
+    }
+    Ok(u32::try_from(n)
+        .unwrap_or(MAX_FIND_LIMIT)
+        .min(MAX_FIND_LIMIT))
+}
+
+// ----------------------------------------------------------------------------
+// Tests — error messages are part of the SDK contract once scripts
+// depend on them; the snapshot-style asserts pin the exact strings.
+// ----------------------------------------------------------------------------
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use serde_json::json;
+
+    fn parse(v: Value) -> Result<DocsFilter, FilterParseError> {
+        parse_filter(&v)
+    }
+
+    #[test]
+    fn empty_object_has_no_conditions() {
+        let f = parse(json!({})).unwrap();
+        assert!(f.conditions.is_empty());
+        assert!(f.sort.is_none());
+        assert!(f.limit.is_none());
+    }
+
+    #[test]
+    fn single_equality_top_level() {
+        let f = parse(json!({ "tier": "gold" })).unwrap();
+        assert_eq!(f.conditions.len(), 1);
+        assert_eq!(f.conditions[0].path.segments(), &["tier".to_string()]);
+        assert_eq!(f.conditions[0].op, ComparisonOp::Eq);
+        assert_eq!(f.conditions[0].value, json!("gold"));
+    }
+
+    #[test]
+    fn multi_field_equality_is_conjunctive() {
+        let f = parse(json!({ "tier": "gold", "status": "active" })).unwrap();
+        assert_eq!(f.conditions.len(), 2);
+    }
+
+    #[test]
+    fn nested_dotted_path() {
+        let f = parse(json!({ "user.email": "a@b" })).unwrap();
+        let cond = &f.conditions[0];
+        assert_eq!(
+            cond.path.segments(),
+            &["user".to_string(), "email".to_string()]
+        );
+    }
+
+    #[test]
+    fn depth_limit_rejects_six_segments() {
+        let err = parse(json!({ "a.b.c.d.e.f": "x" })).unwrap_err();
+        let msg = err.to_string();
+        assert!(msg.contains("exceeds max depth"), "msg: {msg}");
+        assert!(msg.contains('5'), "msg: {msg}");
+    }
+
+    #[test]
+    fn double_dot_rejected() {
+        let err = parse(json!({ "a..b": "x" })).unwrap_err();
+        assert!(err.to_string().contains("empty segment"));
+    }
+
+    #[test]
+    fn leading_dot_rejected() {
+        let err = parse(json!({ ".a": "x" })).unwrap_err();
+        assert!(err.to_string().contains("empty segment"));
+    }
+
+    #[test]
+    fn trailing_dot_rejected() {
+        let err = parse(json!({ "a.": "x" })).unwrap_err();
+        assert!(err.to_string().contains("empty segment"));
+    }
+
+    #[test]
+    fn dollar_prefix_in_path_segment_rejected() {
+        // (The top-level $foo would route to operator dispatch; this
+        // tests deeper segments which should never start with $.)
+        let err = parse(json!({ "x.$inner": "v" })).unwrap_err();
+        assert!(err.to_string().contains("must not start with '$'"));
+    }
+
+    #[test]
+    fn each_supported_operator_parses() {
+        for (key, expected_op) in [
+            ("$eq", ComparisonOp::Eq),
+            ("$ne", ComparisonOp::Ne),
+            ("$gt", ComparisonOp::Gt),
+            ("$gte", ComparisonOp::Gte),
+            ("$lt", ComparisonOp::Lt),
+            ("$lte", ComparisonOp::Lte),
+        ] {
+            let v = json!({ "field": { key: "v" } });
+            let f = parse(v).unwrap();
+            assert_eq!(f.conditions[0].op, expected_op, "key {key}");
+        }
+        // $in needs an array.
+        let f = parse(json!({ "tier": { "$in": ["gold", "platinum"] } })).unwrap();
+        assert_eq!(f.conditions[0].op, ComparisonOp::In);
+    }
+
+    #[test]
+    fn dollar_in_with_non_array_value_rejected() {
+        let err = parse(json!({ "tier": { "$in": "gold" } })).unwrap_err();
+        assert!(err.to_string().contains("'$in'"));
+        assert!(err.to_string().contains("array"));
+    }
+
+    #[test]
+    fn scalar_op_with_object_value_rejected() {
+        let err = parse(json!({ "tier": { "$gt": { "nested": true } } })).unwrap_err();
+        assert!(err.to_string().contains("'$gt'"));
+        assert!(err.to_string().contains("scalar"));
+    }
+
+    /// Snapshot: the v1.2-deferred operator error string is part of
+    /// the SDK contract. Don't change it without a major-version bump.
+    #[test]
+    fn unsupported_operator_message_pins_v1_2_pointer() {
+        let err = parse(json!({ "name": { "$regex": "^A" } })).unwrap_err();
+        assert_eq!(
+            err.to_string(),
+            "docs::find: operator '$regex' is not supported in v1.1.2; planned for v1.2 advanced query"
+        );
+    }
+
+    #[test]
+    fn unsupported_top_level_modifier_rejected() {
+        let err = parse(json!({ "$or": [{ "x": 1 }] })).unwrap_err();
+        assert!(err.to_string().contains("'$or'"));
+        assert!(err.to_string().contains("v1.2"));
+    }
+
+    /// Snapshot: depth-limit error string. Pinned per the SDK contract.
+    #[test]
+    fn depth_limit_message_pinned() {
+        let err = parse(json!({ "a.b.c.d.e.f": 1 })).unwrap_err();
+        assert_eq!(
+            err.to_string(),
+            "docs::find: field path 'a.b.c.d.e.f' exceeds max depth 5"
+        );
+    }
+
+    #[test]
+    fn mixed_shape_operator_object_rejected() {
+        // Object value where some keys are $-prefixed and some aren't
+        // — treated as user data + invalid (the user almost certainly
+        // meant an operator object).
+        let err = parse(json!({ "x": { "$gt": 1, "other": 2 } })).unwrap_err();
+        assert!(err
+            .to_string()
+            .contains("scalar (implicit equality) or an operator map"));
+    }
+
+    #[test]
+    fn sort_asc_and_desc_parse() {
+        let f = parse(json!({ "$sort": { "created_at": 1 } })).unwrap();
+        let sort = f.sort.unwrap();
+        assert_eq!(sort.direction, SortDir::Asc);
+        assert_eq!(sort.path.segments(), &["created_at".to_string()]);
+
+        let f = parse(json!({ "$sort": { "created_at": -1 } })).unwrap();
+        assert_eq!(f.sort.unwrap().direction, SortDir::Desc);
+    }
+
+    #[test]
+    fn sort_with_bad_direction_rejected() {
+        let err = parse(json!({ "$sort": { "x": 2 } })).unwrap_err();
+        assert!(err.to_string().contains("1 (ascending)"));
+    }
+
+    /// Snapshot: multi-field sort error string. Pinned.
+    #[test]
+    fn multi_field_sort_rejected_with_v1_2_pointer() {
+        let err = parse(json!({ "$sort": { "a": 1, "b": -1 } })).unwrap_err();
+        assert_eq!(
+            err.to_string(),
+            "docs::find: multi-field '$sort' is not supported in v1.1.2; planned for v1.2 advanced query"
+        );
+    }
+
+    #[test]
+    fn limit_accepts_non_negative_integer() {
+        let f = parse(json!({ "$limit": 50 })).unwrap();
+        assert_eq!(f.limit, Some(50));
+    }
+
+    #[test]
+    fn limit_clamps_to_max() {
+        let f = parse(json!({ "$limit": 10_000 })).unwrap();
+        assert_eq!(f.limit, Some(MAX_FIND_LIMIT));
+    }
+
+    #[test]
+    fn limit_rejects_negative() {
+        let err = parse(json!({ "$limit": -1 })).unwrap_err();
+        assert!(err.to_string().contains("non-negative"));
+    }
+
+    #[test]
+    fn limit_rejects_non_integer() {
+        let err = parse(json!({ "$limit": "twenty" })).unwrap_err();
+        assert!(err.to_string().contains("integer"));
+    }
+
+    #[test]
+    fn non_object_filter_rejected() {
+        let err = parse(json!("not a map")).unwrap_err();
+        assert!(err.to_string().contains("filter must be a map/object"));
+    }
+
+    #[test]
+    fn dollar_eq_value_can_be_null() {
+        // $ne against null is a valid query (returns docs where field
+        // exists and is not null OR is missing) — so null must be an
+        // accepted scalar.
+        let f = parse(json!({ "deleted_at": { "$ne": null } })).unwrap();
+        assert_eq!(f.conditions[0].op, ComparisonOp::Ne);
+        assert_eq!(f.conditions[0].value, Value::Null);
+    }
+
+    #[test]
+    fn implicit_equality_with_array_value_accepts() {
+        // `{ "tags": ["a", "b"] }` is implicit equality against the
+        // literal array shape. The Postgres query will compare the
+        // text encoding under JSONB; this is valid v1.1.2.
+        let f = parse(json!({ "tags": ["a", "b"] })).unwrap();
+        assert_eq!(f.conditions[0].op, ComparisonOp::Eq);
+    }
+}

crates/manager-core/src/docs_repo.rs

@@ -0,0 +1,556 @@
+//! Low-level Postgres CRUD + filter-query builder over the `docs`
+//! table (migration 0013). Stays storage-only; authorization, event
+//! emission, and empty-collection validation live one layer up in
+//! `DocsServiceImpl`.
+//!
+//! The `find` SQL builder is the security-critical surface. **Every
+//! field-path segment and every comparison value is bound as a
+//! `$N` parameter — never interpolated into the SQL string.** The base
+//! `WHERE app_id = $1 AND collection = $2` clause is fixed and
+//! prepended to every query so cross-app isolation can't be widened by
+//! any operator. See `sql_starts_with_app_collection_predicate`
+//! assertion in tests for the load-bearing guarantee.
+
+use async_trait::async_trait;
+use base64::engine::general_purpose::URL_SAFE_NO_PAD;
+use base64::Engine as _;
+use chrono::{DateTime, Utc};
+use picloud_shared::{AppId, DocId, DocRow, DocsListPage};
+use serde_json::Value;
+use sqlx::postgres::PgRow;
+use sqlx::{PgPool, Postgres, QueryBuilder, Row};
+use uuid::Uuid;
+
+use crate::docs_filter::{ComparisonOp, DocsFilter, SortDir};
+
+#[derive(Debug, thiserror::Error)]
+pub enum DocsRepoError {
+    #[error("database error: {0}")]
+    Db(#[from] sqlx::Error),
+
+    #[error("invalid pagination cursor")]
+    InvalidCursor,
+}
+
+/// Repo surface. The trait is exposed so the service unit tests can
+/// substitute an in-memory backing without spinning up Postgres.
+#[async_trait]
+pub trait DocsRepo: Send + Sync {
+    /// Create a new doc with a server-generated UUID. Returns the
+    /// fully-materialised `DocRow` so the caller has timestamps too
+    /// (no separate select-back round-trip).
+    async fn create(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        data: Value,
+    ) -> Result<DocRow, DocsRepoError>;
+
+    async fn get(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: DocId,
+    ) -> Result<Option<DocRow>, DocsRepoError>;
+
+    /// Filter-based query. The parsed `DocsFilter` ensures every
+    /// field-path segment and operator value is bound as a parameter.
+    async fn find(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        filter: &DocsFilter,
+    ) -> Result<Vec<DocRow>, DocsRepoError>;
+
+    /// Full document replace. Returns `Some(previous_data)` on
+    /// success, `None` if no doc matched (the service maps that to
+    /// `DocsError::NotFound`). The prev value is the input to the
+    /// emitted update event's `old_payload`.
+    async fn update(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: DocId,
+        data: Value,
+    ) -> Result<Option<Value>, DocsRepoError>;
+
+    /// Returns the deleted doc's data if it existed, `None` if no
+    /// such doc. The caller converts `Some` → `Ok(true)` for the SDK's
+    /// was-present return; the `Value` feeds the delete event's
+    /// `old_payload`.
+    async fn delete(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: DocId,
+    ) -> Result<Option<Value>, DocsRepoError>;
+
+    async fn list(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        cursor: Option<&str>,
+        limit: u32,
+    ) -> Result<DocsListPage, DocsRepoError>;
+}
+
+pub struct PostgresDocsRepo {
+    pool: PgPool,
+}
+
+impl PostgresDocsRepo {
+    #[must_use]
+    pub fn new(pool: PgPool) -> Self {
+        Self { pool }
+    }
+}
+
+/// Hard ceiling on `list` page size — mirrors KV's `KV_LIST_MAX_LIMIT`.
+/// Scripts that pass anything larger get silently clamped.
+const DOCS_LIST_MAX_LIMIT: u32 = 1_000;
+const DOCS_LIST_DEFAULT_LIMIT: u32 = 100;
+
+#[async_trait]
+impl DocsRepo for PostgresDocsRepo {
+    async fn create(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        data: Value,
+    ) -> Result<DocRow, DocsRepoError> {
+        let id = Uuid::new_v4();
+        let row: (DateTime<Utc>, DateTime<Utc>) = sqlx::query_as(
+            "INSERT INTO docs (app_id, collection, id, data) \
+             VALUES ($1, $2, $3, $4) \
+             RETURNING created_at, updated_at",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(id)
+        .bind(&data)
+        .fetch_one(&self.pool)
+        .await?;
+        Ok(DocRow {
+            id,
+            data,
+            created_at: row.0,
+            updated_at: row.1,
+        })
+    }
+
+    async fn get(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: DocId,
+    ) -> Result<Option<DocRow>, DocsRepoError> {
+        let row: Option<(Value, DateTime<Utc>, DateTime<Utc>)> = sqlx::query_as(
+            "SELECT data, created_at, updated_at FROM docs \
+             WHERE app_id = $1 AND collection = $2 AND id = $3",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(id)
+        .fetch_optional(&self.pool)
+        .await?;
+        Ok(row.map(|(data, created_at, updated_at)| DocRow {
+            id,
+            data,
+            created_at,
+            updated_at,
+        }))
+    }
+
+    async fn find(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        filter: &DocsFilter,
+    ) -> Result<Vec<DocRow>, DocsRepoError> {
+        let mut qb = build_find_query(app_id, collection, filter);
+        let rows = qb.build().fetch_all(&self.pool).await?;
+        rows.into_iter().map(row_to_doc).collect()
+    }
+
+    async fn update(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: DocId,
+        data: Value,
+    ) -> Result<Option<Value>, DocsRepoError> {
+        // Same CTE shape as KV's set ([kv_repo.rs:101-132]): SELECT the
+        // previous data before the UPDATE so the service can emit
+        // `prev_data` in the update ServiceEvent. Single statement, no
+        // explicit transaction. Inherits KV's last-writer-wins race
+        // under concurrent writers; documented as a known limitation
+        // for v1.1.2.
+        let row: Option<(Option<Value>,)> = sqlx::query_as(
+            "WITH prev AS ( \
+                SELECT data FROM docs \
+                WHERE app_id = $1 AND collection = $2 AND id = $3 \
+             ), \
+             updated AS ( \
+                UPDATE docs SET data = $4, updated_at = NOW() \
+                WHERE app_id = $1 AND collection = $2 AND id = $3 \
+                RETURNING 1 \
+             ) \
+             SELECT (SELECT data FROM prev) FROM updated",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(id)
+        .bind(&data)
+        .fetch_optional(&self.pool)
+        .await?;
+        // `row` is None when the UPDATE matched no rows (missing doc);
+        // Some((Some(prev),)) on success. `data` is JSONB NOT NULL so
+        // the inner Option is always Some when prev exists.
+        Ok(row.and_then(|(v,)| v))
+    }
+
+    async fn delete(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: DocId,
+    ) -> Result<Option<Value>, DocsRepoError> {
+        let row: Option<(Value,)> = sqlx::query_as(
+            "DELETE FROM docs \
+             WHERE app_id = $1 AND collection = $2 AND id = $3 \
+             RETURNING data",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(id)
+        .fetch_optional(&self.pool)
+        .await?;
+        Ok(row.map(|(v,)| v))
+    }
+
+    async fn list(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        cursor: Option<&str>,
+        limit: u32,
+    ) -> Result<DocsListPage, DocsRepoError> {
+        let limit = if limit == 0 {
+            DOCS_LIST_DEFAULT_LIMIT
+        } else {
+            limit.min(DOCS_LIST_MAX_LIMIT)
+        };
+
+        let last_id = match cursor {
+            Some(c) => Some(decode_cursor(c)?),
+            None => None,
+        };
+
+        let take = i64::from(limit) + 1;
+        let rows: Vec<(Uuid, Value, DateTime<Utc>, DateTime<Utc>)> = sqlx::query_as(
+            "SELECT id, data, created_at, updated_at FROM docs \
+             WHERE app_id = $1 AND collection = $2 \
+               AND ($3::uuid IS NULL OR id > $3) \
+             ORDER BY id ASC \
+             LIMIT $4",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(last_id)
+        .bind(take)
+        .fetch_all(&self.pool)
+        .await?;
+
+        let mut docs: Vec<DocRow> = rows
+            .into_iter()
+            .map(|(id, data, created_at, updated_at)| DocRow {
+                id,
+                data,
+                created_at,
+                updated_at,
+            })
+            .collect();
+        let next_cursor = if docs.len() > limit as usize {
+            docs.truncate(limit as usize);
+            docs.last().map(|d| encode_cursor(&d.id))
+        } else {
+            None
+        };
+
+        Ok(DocsListPage { docs, next_cursor })
+    }
+}
+
+fn row_to_doc(row: PgRow) -> Result<DocRow, DocsRepoError> {
+    Ok(DocRow {
+        id: row.try_get("id")?,
+        data: row.try_get("data")?,
+        created_at: row.try_get("created_at")?,
+        updated_at: row.try_get("updated_at")?,
+    })
+}
+
+fn encode_cursor(last_id: &Uuid) -> String {
+    URL_SAFE_NO_PAD.encode(last_id.as_bytes())
+}
+
+fn decode_cursor(cursor: &str) -> Result<Uuid, DocsRepoError> {
+    let bytes = URL_SAFE_NO_PAD
+        .decode(cursor)
+        .map_err(|_| DocsRepoError::InvalidCursor)?;
+    let arr: [u8; 16] = bytes
+        .as_slice()
+        .try_into()
+        .map_err(|_| DocsRepoError::InvalidCursor)?;
+    Ok(Uuid::from_bytes(arr))
+}
+
+// ----------------------------------------------------------------------------
+// SQL builder — the load-bearing security surface.
+//
+// Every field-path segment + every comparison value goes through
+// `QueryBuilder::push_bind`, which appends `$N` to the SQL string and
+// binds the value as a parameter. The only literal strings appended to
+// the SQL are: hardcoded SQL fragments (SELECT/WHERE/AND/etc.) and
+// hardcoded operator strings ("=", "IS DISTINCT FROM", ">", "ASC", …).
+// **No user input ever lands in the SQL text unparameterized.**
+// ----------------------------------------------------------------------------
+
+fn build_find_query<'a>(
+    app_id: AppId,
+    collection: &'a str,
+    filter: &'a DocsFilter,
+) -> QueryBuilder<'a, Postgres> {
+    let mut qb =
+        QueryBuilder::new("SELECT id, data, created_at, updated_at FROM docs WHERE app_id = ");
+    qb.push_bind(app_id.into_inner());
+    qb.push(" AND collection = ");
+    qb.push_bind(collection);
+
+    for cond in &filter.conditions {
+        qb.push(" AND ");
+        emit_condition(&mut qb, cond);
+    }
+
+    qb.push(" ORDER BY ");
+    if let Some(sort) = &filter.sort {
+        push_jsonb_path(&mut qb, sort.path.segments());
+        qb.push(match sort.direction {
+            SortDir::Asc => " ASC",
+            SortDir::Desc => " DESC",
+        });
+        qb.push(", id ASC");
+    } else {
+        qb.push("id ASC");
+    }
+
+    let limit = filter
+        .limit
+        .map_or(DOCS_LIST_MAX_LIMIT, |l| l.min(DOCS_LIST_MAX_LIMIT));
+    qb.push(" LIMIT ");
+    qb.push_bind(i64::from(limit));
+
+    qb
+}
+
+fn emit_condition<'a>(
+    qb: &mut QueryBuilder<'a, Postgres>,
+    cond: &'a crate::docs_filter::FieldCondition,
+) {
+    push_jsonb_path(qb, cond.path.segments());
+    match cond.op {
+        ComparisonOp::Eq => {
+            if cond.value.is_null() {
+                qb.push(" IS NULL");
+            } else {
+                qb.push(" = ");
+                qb.push_bind(value_to_text(&cond.value));
+            }
+        }
+        ComparisonOp::Ne => {
+            // IS DISTINCT FROM correctly handles NULL on either side
+            // (would otherwise silently exclude rows with missing
+            // paths). Holds for the literal-NULL case too.
+            if cond.value.is_null() {
+                qb.push(" IS NOT NULL");
+            } else {
+                qb.push(" IS DISTINCT FROM ");
+                qb.push_bind(value_to_text(&cond.value));
+            }
+        }
+        ComparisonOp::Gt => {
+            qb.push(" > ");
+            qb.push_bind(value_to_text(&cond.value));
+        }
+        ComparisonOp::Gte => {
+            qb.push(" >= ");
+            qb.push_bind(value_to_text(&cond.value));
+        }
+        ComparisonOp::Lt => {
+            qb.push(" < ");
+            qb.push_bind(value_to_text(&cond.value));
+        }
+        ComparisonOp::Lte => {
+            qb.push(" <= ");
+            qb.push_bind(value_to_text(&cond.value));
+        }
+        ComparisonOp::In => {
+            qb.push(" = ANY(");
+            let texts: Vec<Option<String>> = cond
+                .value
+                .as_array()
+                .map(|arr| arr.iter().map(value_to_text).collect())
+                .unwrap_or_default();
+            qb.push_bind(texts);
+            qb.push(")");
+        }
+    }
+}
+
+/// Append `jsonb_extract_path_text(data, $N1, $N2, …)` with each
+/// segment bound as a separate text parameter. Variadic path lengths
+/// (1–5) all flow through this single helper.
+fn push_jsonb_path<'a>(qb: &mut QueryBuilder<'a, Postgres>, segments: &'a [String]) {
+    qb.push("jsonb_extract_path_text(data");
+    for seg in segments {
+        qb.push(", ");
+        qb.push_bind(seg.as_str());
+    }
+    qb.push(")");
+}
+
+/// JSON scalar → TEXT for binding. `Value::Null` is preserved as
+/// `None` so the binding lands as SQL NULL (handled specially above for
+/// `Eq` / `Ne`). Arrays + objects serialize to compact JSON; the user
+/// is comparing against the JSONB text rendering, which is consistent
+/// with `jsonb_extract_path_text`'s output for those types.
+fn value_to_text(v: &Value) -> Option<String> {
+    match v {
+        Value::Null => None,
+        Value::String(s) => Some(s.clone()),
+        Value::Bool(b) => Some(b.to_string()),
+        Value::Number(n) => Some(n.to_string()),
+        Value::Array(_) | Value::Object(_) => Some(v.to_string()),
+    }
+}
+
+// ----------------------------------------------------------------------------
+// SQL-shape guardrail tests — pure (no DB) so they run in the default
+// test suite. These are the highest-stakes tests in the release: they
+// pin the cross-app isolation invariant at the SQL level.
+// ----------------------------------------------------------------------------
+
+#[cfg(test)]
+mod sql_shape_tests {
+    use super::*;
+    use crate::docs_filter::parse_filter;
+    use serde_json::json;
+
+    fn sql_for(filter_json: serde_json::Value) -> String {
+        let filter = parse_filter(&filter_json).unwrap();
+        let qb = build_find_query(AppId::new(), "users", &filter);
+        qb.sql().to_string()
+    }
+
+    /// **Load-bearing**: every generated SELECT begins
+    /// `WHERE app_id = $1 AND collection = $2`. The app_id parameter
+    /// is the cross-app isolation gate. No user-supplied filter
+    /// fragment can ever appear before these clauses.
+    #[test]
+    fn every_query_starts_with_app_id_and_collection_predicate() {
+        let cases = vec![
+            json!({}),
+            json!({ "tier": "gold" }),
+            json!({ "created_at": { "$gt": "2026-01-01" } }),
+            json!({ "tier": { "$in": ["gold", "platinum"] } }),
+            json!({ "tier": "gold", "status": "active" }),
+            json!({ "$sort": { "created_at": -1 }, "$limit": 5 }),
+            json!({ "tier": "gold", "$sort": { "created_at": 1 } }),
+            json!({ "deleted_at": { "$ne": null } }),
+        ];
+        for case in cases {
+            let sql = sql_for(case.clone());
+            assert!(
+                sql.starts_with(
+                    "SELECT id, data, created_at, updated_at FROM docs WHERE app_id = $1 AND collection = $2"
+                ),
+                "filter {case} produced SQL: {sql}"
+            );
+        }
+    }
+
+    /// Every comparison value lands as a `$N` placeholder — there
+    /// should be NO double-quoted string literal in the SQL after the
+    /// fixed prefix. (This guards against an accidental `format!`
+    /// regression.)
+    #[test]
+    fn no_user_string_literal_in_sql() {
+        let sql = sql_for(json!({ "tier": "gold; DROP TABLE docs;--" }));
+        assert!(!sql.contains("gold"), "value leaked into SQL string: {sql}");
+        assert!(!sql.contains("DROP"), "value leaked into SQL string: {sql}");
+    }
+
+    /// Field-path segments also bind as parameters. A user passing a
+    /// path that looks like SQL keywords doesn't change the structure.
+    #[test]
+    fn no_user_path_literal_in_sql() {
+        let sql = sql_for(json!({ "drop_table_users": "v" }));
+        assert!(
+            !sql.contains("drop_table_users"),
+            "path leaked into SQL string: {sql}"
+        );
+    }
+
+    #[test]
+    fn empty_filter_sql_has_no_extra_conditions() {
+        let sql = sql_for(json!({}));
+        // After the fixed prefix, only ORDER BY + LIMIT — no `AND`s.
+        let suffix = sql
+            .trim_start_matches(
+                "SELECT id, data, created_at, updated_at FROM docs WHERE app_id = $1 AND collection = $2",
+            )
+            .trim();
+        assert!(
+            suffix.starts_with("ORDER BY"),
+            "expected ORDER BY immediately after base WHERE; got: {suffix}"
+        );
+    }
+
+    #[test]
+    fn eq_with_null_emits_is_null() {
+        let sql = sql_for(json!({ "x": null }));
+        assert!(sql.contains("IS NULL"), "sql: {sql}");
+    }
+
+    #[test]
+    fn ne_with_null_emits_is_not_null() {
+        let sql = sql_for(json!({ "x": { "$ne": null } }));
+        assert!(sql.contains("IS NOT NULL"), "sql: {sql}");
+    }
+
+    #[test]
+    fn ne_with_value_uses_is_distinct_from() {
+        // IS DISTINCT FROM, NOT <> — see ComparisonOp::Ne comment.
+        let sql = sql_for(json!({ "x": { "$ne": "v" } }));
+        assert!(sql.contains("IS DISTINCT FROM"), "sql: {sql}");
+        assert!(!sql.contains(" <> "), "sql: {sql}");
+    }
+
+    #[test]
+    fn in_emits_any_array() {
+        let sql = sql_for(json!({ "x": { "$in": ["a", "b"] } }));
+        assert!(sql.contains("= ANY"), "sql: {sql}");
+    }
+
+    #[test]
+    fn sort_appends_tiebreaker_id_asc() {
+        let sql = sql_for(json!({ "$sort": { "created_at": -1 } }));
+        assert!(sql.contains("DESC, id ASC"), "sql: {sql}");
+    }
+
+    #[test]
+    fn jsonb_extract_path_used_for_field_access() {
+        let sql = sql_for(json!({ "user.email": "a@b" }));
+        assert!(sql.contains("jsonb_extract_path_text(data"), "sql: {sql}");
+    }
+}

crates/manager-core/src/docs_service.rs

@@ -0,0 +1,892 @@
+//! `DocsServiceImpl` — wires the `DocsRepo` underneath the
+//! `picloud_shared::DocsService` trait that scripts see via the Rhai
+//! bridge.
+//!
+//! Layers added here (vs the raw repo):
+//!
+//! 1. Empty-collection rejection at the SDK boundary
+//!    (`docs/sdk-shape.md`).
+//! 2. `data` must be a JSON object for create + update. (The repo
+//!    accepts anything serde_json can serialise; the SDK contract
+//!    pins documents to map shape so dotted-path queries make sense.)
+//! 3. **Script-as-gate authz**: when `cx.principal.is_some()` we run
+//!    `authz::require(...)`; when it's `None` (public unauthenticated
+//!    HTTP — the common case for public routes) we skip the check.
+//!    Cross-app isolation isn't affected — every query is keyed by
+//!    `cx.app_id`, never an argument.
+//! 4. Query DSL parse — `find`/`find_one` parse the opaque filter
+//!    into `DocsFilter` before passing it down. Parse errors map to
+//!    `DocsError::InvalidFilter` / `UnsupportedOperator` with the
+//!    parser's message verbatim (script-visible).
+//! 5. `ServiceEvent` emission after each mutation (`create` / `update`
+//!    / `delete`). The outbox emitter (when wired) turns these into
+//!    docs-trigger fan-out via `OutboxEventEmitter::emit_docs`.
+
+use std::sync::Arc;
+
+use async_trait::async_trait;
+use picloud_shared::{
+    DocId, DocRow, DocsError, DocsListPage, DocsService, SdkCallCx, ServiceEvent,
+    ServiceEventEmitter,
+};
+
+use crate::authz::{self, AuthzRepo, Capability};
+use crate::docs_filter::{parse_filter, FilterParseError};
+use crate::docs_repo::{DocsRepo, DocsRepoError};
+
+pub struct DocsServiceImpl {
+    repo: Arc<dyn DocsRepo>,
+    authz: Arc<dyn AuthzRepo>,
+    events: Arc<dyn ServiceEventEmitter>,
+}
+
+impl DocsServiceImpl {
+    #[must_use]
+    pub fn new(
+        repo: Arc<dyn DocsRepo>,
+        authz: Arc<dyn AuthzRepo>,
+        events: Arc<dyn ServiceEventEmitter>,
+    ) -> Self {
+        Self {
+            repo,
+            authz,
+            events,
+        }
+    }
+
+    async fn check_read(&self, cx: &SdkCallCx) -> Result<(), DocsError> {
+        if let Some(ref principal) = cx.principal {
+            authz::require(&*self.authz, principal, Capability::AppDocsRead(cx.app_id))
+                .await
+                .map_err(|_| DocsError::Forbidden)?;
+        }
+        Ok(())
+    }
+
+    async fn check_write(&self, cx: &SdkCallCx) -> Result<(), DocsError> {
+        if let Some(ref principal) = cx.principal {
+            authz::require(&*self.authz, principal, Capability::AppDocsWrite(cx.app_id))
+                .await
+                .map_err(|_| DocsError::Forbidden)?;
+        }
+        Ok(())
+    }
+}
+
+fn validate_collection(collection: &str) -> Result<(), DocsError> {
+    if collection.is_empty() {
+        return Err(DocsError::InvalidCollection);
+    }
+    Ok(())
+}
+
+fn validate_data(data: &serde_json::Value) -> Result<(), DocsError> {
+    if !data.is_object() {
+        return Err(DocsError::InvalidData);
+    }
+    Ok(())
+}
+
+impl From<DocsRepoError> for DocsError {
+    fn from(e: DocsRepoError) -> Self {
+        Self::Backend(e.to_string())
+    }
+}
+
+impl From<FilterParseError> for DocsError {
+    fn from(e: FilterParseError) -> Self {
+        match e {
+            FilterParseError::InvalidFilter(s) => Self::InvalidFilter(s),
+            FilterParseError::UnsupportedOperator(s) => Self::UnsupportedOperator(s),
+        }
+    }
+}
+
+#[async_trait]
+impl DocsService for DocsServiceImpl {
+    async fn create(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        data: serde_json::Value,
+    ) -> Result<DocId, DocsError> {
+        validate_collection(collection)?;
+        validate_data(&data)?;
+        self.check_write(cx).await?;
+        let row = self
+            .repo
+            .create(cx.app_id, collection, data.clone())
+            .await?;
+        // Best-effort emit — a failed emit logs but does not roll back
+        // the write (mirrors KV's pattern).
+        if let Err(e) = self
+            .events
+            .emit(
+                cx,
+                ServiceEvent {
+                    source: "docs",
+                    op: "create",
+                    collection: Some(collection.to_string()),
+                    key: Some(row.id.to_string()),
+                    payload: Some(data),
+                    old_payload: None,
+                },
+            )
+            .await
+        {
+            tracing::warn!(error = %e, source = "docs", op = "create", "event emit failed");
+        }
+        Ok(row.id)
+    }
+
+    async fn get(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: DocId,
+    ) -> Result<Option<DocRow>, DocsError> {
+        validate_collection(collection)?;
+        self.check_read(cx).await?;
+        Ok(self.repo.get(cx.app_id, collection, id).await?)
+    }
+
+    async fn find(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        filter: serde_json::Value,
+    ) -> Result<Vec<DocRow>, DocsError> {
+        validate_collection(collection)?;
+        self.check_read(cx).await?;
+        let parsed = parse_filter(&filter)?;
+        Ok(self.repo.find(cx.app_id, collection, &parsed).await?)
+    }
+
+    async fn find_one(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        filter: serde_json::Value,
+    ) -> Result<Option<DocRow>, DocsError> {
+        validate_collection(collection)?;
+        self.check_read(cx).await?;
+        let mut parsed = parse_filter(&filter)?;
+        // Inject the implicit `LIMIT 1` for find_one — explicit
+        // caller-supplied `$limit` wins.
+        if parsed.limit.is_none() {
+            parsed.limit = Some(1);
+        }
+        let rows = self.repo.find(cx.app_id, collection, &parsed).await?;
+        Ok(rows.into_iter().next())
+    }
+
+    async fn update(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: DocId,
+        data: serde_json::Value,
+    ) -> Result<(), DocsError> {
+        validate_collection(collection)?;
+        validate_data(&data)?;
+        self.check_write(cx).await?;
+        let previous = self
+            .repo
+            .update(cx.app_id, collection, id, data.clone())
+            .await?;
+        match previous {
+            Some(prev) => {
+                if let Err(e) = self
+                    .events
+                    .emit(
+                        cx,
+                        ServiceEvent {
+                            source: "docs",
+                            op: "update",
+                            collection: Some(collection.to_string()),
+                            key: Some(id.to_string()),
+                            payload: Some(data),
+                            old_payload: Some(prev),
+                        },
+                    )
+                    .await
+                {
+                    tracing::warn!(error = %e, source = "docs", op = "update", "event emit failed");
+                }
+                Ok(())
+            }
+            None => Err(DocsError::NotFound),
+        }
+    }
+
+    async fn delete(&self, cx: &SdkCallCx, collection: &str, id: DocId) -> Result<bool, DocsError> {
+        validate_collection(collection)?;
+        self.check_write(cx).await?;
+        let previous = self.repo.delete(cx.app_id, collection, id).await?;
+        let was_present = previous.is_some();
+        if let Some(prev) = previous {
+            if let Err(e) = self
+                .events
+                .emit(
+                    cx,
+                    ServiceEvent {
+                        source: "docs",
+                        op: "delete",
+                        collection: Some(collection.to_string()),
+                        key: Some(id.to_string()),
+                        payload: None,
+                        old_payload: Some(prev),
+                    },
+                )
+                .await
+            {
+                tracing::warn!(error = %e, source = "docs", op = "delete", "event emit failed");
+            }
+        }
+        Ok(was_present)
+    }
+
+    async fn list(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        cursor: Option<&str>,
+        limit: u32,
+    ) -> Result<DocsListPage, DocsError> {
+        validate_collection(collection)?;
+        self.check_read(cx).await?;
+        Ok(self.repo.list(cx.app_id, collection, cursor, limit).await?)
+    }
+}
+
+// ----------------------------------------------------------------------------
+// Tests — in-memory DocsRepo so unit tests don't need Postgres.
+// ----------------------------------------------------------------------------
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::authz::{AuthzError, AuthzRepo};
+    use crate::docs_filter::DocsFilter;
+    use async_trait::async_trait;
+    use chrono::Utc;
+    use picloud_shared::{
+        AdminUserId, AppId, AppRole, ExecutionId, InstanceRole, NoopEventEmitter, Principal,
+        RequestId, ScriptId, UserId,
+    };
+    use serde_json::json;
+    use std::collections::BTreeMap;
+    use std::sync::Arc;
+    use tokio::sync::Mutex;
+    use uuid::Uuid;
+
+    /// In-memory backing: BTreeMap keyed by `(app_id, collection, id)`
+    /// so iteration is naturally ordered for stable cursor pagination
+    /// (matches the Postgres `ORDER BY id ASC`).
+    #[derive(Default)]
+    struct InMemoryDocsRepo {
+        data: Mutex<BTreeMap<(AppId, String, DocId), DocRow>>,
+    }
+
+    #[async_trait]
+    impl DocsRepo for InMemoryDocsRepo {
+        async fn create(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            data: serde_json::Value,
+        ) -> Result<DocRow, DocsRepoError> {
+            let id = Uuid::new_v4();
+            let now = Utc::now();
+            let row = DocRow {
+                id,
+                data,
+                created_at: now,
+                updated_at: now,
+            };
+            self.data
+                .lock()
+                .await
+                .insert((app_id, collection.to_string(), id), row.clone());
+            Ok(row)
+        }
+
+        async fn get(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            id: DocId,
+        ) -> Result<Option<DocRow>, DocsRepoError> {
+            Ok(self
+                .data
+                .lock()
+                .await
+                .get(&(app_id, collection.to_string(), id))
+                .cloned())
+        }
+
+        async fn find(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            filter: &DocsFilter,
+        ) -> Result<Vec<DocRow>, DocsRepoError> {
+            let map = self.data.lock().await;
+            let mut out: Vec<DocRow> = map
+                .iter()
+                .filter(|((a, c, _), _)| *a == app_id && c == collection)
+                .map(|(_, v)| v.clone())
+                .filter(|row| in_memory_matches(row, filter))
+                .collect();
+            if let Some(sort) = &filter.sort {
+                let path = sort.path.segments().to_vec();
+                let dir = sort.direction;
+                out.sort_by(|a, b| {
+                    let av = extract_path_str(&a.data, &path);
+                    let bv = extract_path_str(&b.data, &path);
+                    let ord = av.cmp(&bv);
+                    match dir {
+                        crate::docs_filter::SortDir::Asc => ord,
+                        crate::docs_filter::SortDir::Desc => ord.reverse(),
+                    }
+                });
+            } else {
+                out.sort_by_key(|d| d.id);
+            }
+            if let Some(limit) = filter.limit {
+                out.truncate(limit as usize);
+            }
+            Ok(out)
+        }
+
+        async fn update(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            id: DocId,
+            data: serde_json::Value,
+        ) -> Result<Option<serde_json::Value>, DocsRepoError> {
+            let mut map = self.data.lock().await;
+            let key = (app_id, collection.to_string(), id);
+            let Some(existing) = map.get_mut(&key) else {
+                return Ok(None);
+            };
+            let prev = std::mem::replace(&mut existing.data, data);
+            existing.updated_at = Utc::now();
+            Ok(Some(prev))
+        }
+
+        async fn delete(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            id: DocId,
+        ) -> Result<Option<serde_json::Value>, DocsRepoError> {
+            Ok(self
+                .data
+                .lock()
+                .await
+                .remove(&(app_id, collection.to_string(), id))
+                .map(|row| row.data))
+        }
+
+        async fn list(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            cursor: Option<&str>,
+            limit: u32,
+        ) -> Result<DocsListPage, DocsRepoError> {
+            let map = self.data.lock().await;
+            let last_id = cursor
+                .map(|c| Uuid::parse_str(c).map_err(|_| DocsRepoError::InvalidCursor))
+                .transpose()?;
+            let mut docs: Vec<DocRow> = map
+                .iter()
+                .filter(|((a, c, _), _)| *a == app_id && c == collection)
+                .map(|(_, v)| v.clone())
+                .filter(|d| last_id.is_none_or(|lid| d.id > lid))
+                .collect();
+            docs.sort_by_key(|d| d.id);
+            let take = if limit == 0 {
+                usize::MAX
+            } else {
+                limit as usize
+            };
+            let next_cursor = if docs.len() > take {
+                docs.truncate(take);
+                docs.last().map(|d| d.id.to_string())
+            } else {
+                None
+            };
+            Ok(DocsListPage { docs, next_cursor })
+        }
+    }
+
+    /// Best-effort in-memory filter eval mirroring the Postgres
+    /// semantics: extract each field path as a text-form string, then
+    /// apply the operator. Good enough for the unit tests; production
+    /// always goes through the Postgres impl.
+    fn in_memory_matches(row: &DocRow, filter: &DocsFilter) -> bool {
+        for cond in &filter.conditions {
+            let actual = extract_path_str(&row.data, cond.path.segments());
+            if !cond_matches(actual.as_ref(), cond) {
+                return false;
+            }
+        }
+        true
+    }
+
+    fn cond_matches(actual: Option<&String>, cond: &crate::docs_filter::FieldCondition) -> bool {
+        use crate::docs_filter::ComparisonOp::*;
+        let actual: Option<&str> = actual.map(String::as_str);
+        let want = json_text(&cond.value);
+        let want_ref: Option<&str> = want.as_deref();
+        match cond.op {
+            Eq => actual == want_ref,
+            Ne => actual != want_ref,
+            Gt => actual.zip(want_ref).is_some_and(|(a, b)| a > b),
+            Gte => actual.zip(want_ref).is_some_and(|(a, b)| a >= b),
+            Lt => actual.zip(want_ref).is_some_and(|(a, b)| a < b),
+            Lte => actual.zip(want_ref).is_some_and(|(a, b)| a <= b),
+            In => {
+                let Some(arr) = cond.value.as_array() else {
+                    return false;
+                };
+                arr.iter().any(|v| actual == json_text(v).as_deref())
+            }
+        }
+    }
+
+    fn extract_path_str(value: &serde_json::Value, segments: &[String]) -> Option<String> {
+        let mut cur = value;
+        for seg in segments {
+            cur = cur.as_object()?.get(seg)?;
+        }
+        json_text(cur)
+    }
+
+    fn json_text(v: &serde_json::Value) -> Option<String> {
+        match v {
+            serde_json::Value::Null => None,
+            serde_json::Value::String(s) => Some(s.clone()),
+            serde_json::Value::Bool(b) => Some(b.to_string()),
+            serde_json::Value::Number(n) => Some(n.to_string()),
+            serde_json::Value::Array(_) | serde_json::Value::Object(_) => Some(v.to_string()),
+        }
+    }
+
+    #[derive(Default)]
+    struct DenyingAuthzRepo;
+
+    #[async_trait]
+    impl AuthzRepo for DenyingAuthzRepo {
+        async fn membership(
+            &self,
+            _user_id: UserId,
+            _app_id: AppId,
+        ) -> Result<Option<AppRole>, AuthzError> {
+            Ok(None)
+        }
+    }
+
+    #[derive(Default)]
+    struct AllowingAuthzRepo;
+
+    #[async_trait]
+    impl AuthzRepo for AllowingAuthzRepo {
+        async fn membership(
+            &self,
+            _user_id: UserId,
+            _app_id: AppId,
+        ) -> Result<Option<AppRole>, AuthzError> {
+            Ok(Some(AppRole::Editor))
+        }
+    }
+
+    fn anon_cx(app_id: AppId) -> SdkCallCx {
+        SdkCallCx {
+            app_id,
+            script_id: ScriptId::new(),
+            principal: None,
+            execution_id: ExecutionId::new(),
+            request_id: RequestId::new(),
+            trigger_depth: 0,
+            root_execution_id: ExecutionId::new(),
+            is_dead_letter_handler: false,
+            event: None,
+        }
+    }
+
+    fn owner_cx(app_id: AppId) -> SdkCallCx {
+        SdkCallCx {
+            app_id,
+            script_id: ScriptId::new(),
+            principal: Some(Principal {
+                user_id: AdminUserId::new(),
+                instance_role: InstanceRole::Owner,
+                scopes: None,
+                app_binding: None,
+            }),
+            execution_id: ExecutionId::new(),
+            request_id: RequestId::new(),
+            trigger_depth: 0,
+            root_execution_id: ExecutionId::new(),
+            is_dead_letter_handler: false,
+            event: None,
+        }
+    }
+
+    fn member_no_role_cx(app_id: AppId) -> SdkCallCx {
+        SdkCallCx {
+            app_id,
+            script_id: ScriptId::new(),
+            principal: Some(Principal {
+                user_id: AdminUserId::new(),
+                instance_role: InstanceRole::Member,
+                scopes: None,
+                app_binding: None,
+            }),
+            execution_id: ExecutionId::new(),
+            request_id: RequestId::new(),
+            trigger_depth: 0,
+            root_execution_id: ExecutionId::new(),
+            is_dead_letter_handler: false,
+            event: None,
+        }
+    }
+
+    fn svc() -> DocsServiceImpl {
+        DocsServiceImpl::new(
+            Arc::new(InMemoryDocsRepo::default()),
+            Arc::new(DenyingAuthzRepo),
+            Arc::new(NoopEventEmitter),
+        )
+    }
+
+    fn svc_allowing() -> DocsServiceImpl {
+        DocsServiceImpl::new(
+            Arc::new(InMemoryDocsRepo::default()),
+            Arc::new(AllowingAuthzRepo),
+            Arc::new(NoopEventEmitter),
+        )
+    }
+
+    #[tokio::test]
+    async fn create_then_get_round_trips() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let id = s
+            .create(&cx, "users", json!({ "name": "Alice" }))
+            .await
+            .unwrap();
+        let row = s.get(&cx, "users", id).await.unwrap().unwrap();
+        assert_eq!(row.id, id);
+        assert_eq!(row.data, json!({ "name": "Alice" }));
+    }
+
+    #[tokio::test]
+    async fn get_missing_returns_none() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let v = s.get(&cx, "users", Uuid::new_v4()).await.unwrap();
+        assert!(v.is_none());
+    }
+
+    #[tokio::test]
+    async fn update_missing_returns_not_found() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let err = s
+            .update(&cx, "users", Uuid::new_v4(), json!({ "x": 1 }))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, DocsError::NotFound));
+    }
+
+    #[tokio::test]
+    async fn delete_missing_returns_false() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let was_present = s.delete(&cx, "users", Uuid::new_v4()).await.unwrap();
+        assert!(!was_present);
+    }
+
+    #[tokio::test]
+    async fn delete_present_returns_true() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let id = s.create(&cx, "users", json!({ "x": 1 })).await.unwrap();
+        let was_present = s.delete(&cx, "users", id).await.unwrap();
+        assert!(was_present);
+    }
+
+    #[tokio::test]
+    async fn update_present_succeeds() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let id = s.create(&cx, "users", json!({ "x": 1 })).await.unwrap();
+        s.update(&cx, "users", id, json!({ "x": 2 })).await.unwrap();
+        let row = s.get(&cx, "users", id).await.unwrap().unwrap();
+        assert_eq!(row.data, json!({ "x": 2 }));
+    }
+
+    #[tokio::test]
+    async fn empty_collection_rejected() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let err = s.create(&cx, "", json!({})).await.unwrap_err();
+        assert!(matches!(err, DocsError::InvalidCollection));
+    }
+
+    #[tokio::test]
+    async fn create_with_non_object_data_rejected() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let err = s.create(&cx, "users", json!(42)).await.unwrap_err();
+        assert!(matches!(err, DocsError::InvalidData));
+    }
+
+    #[tokio::test]
+    async fn update_with_non_object_data_rejected() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let id = s.create(&cx, "users", json!({ "x": 1 })).await.unwrap();
+        let err = s
+            .update(&cx, "users", id, json!("not an object"))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, DocsError::InvalidData));
+    }
+
+    /// Load-bearing: a script with `cx.app_id = A` must NOT see
+    /// documents created under `cx.app_id = B`. Cross-app isolation
+    /// boundary; tested through both `get` and `find` because each
+    /// path could conceivably leak independently.
+    #[tokio::test]
+    async fn cross_app_isolation_via_cx_app_id() {
+        let s = svc();
+        let app_a = AppId::new();
+        let app_b = AppId::new();
+        let cx_a = anon_cx(app_a);
+        let cx_b = anon_cx(app_b);
+
+        let id_a = s
+            .create(&cx_a, "shared", json!({ "from": "a" }))
+            .await
+            .unwrap();
+        let id_b = s
+            .create(&cx_b, "shared", json!({ "from": "b" }))
+            .await
+            .unwrap();
+        assert_ne!(id_a, id_b);
+
+        // Each app sees only its own doc via get.
+        assert!(s.get(&cx_a, "shared", id_b).await.unwrap().is_none());
+        assert!(s.get(&cx_b, "shared", id_a).await.unwrap().is_none());
+
+        // And via find.
+        let from_a = s.find(&cx_a, "shared", json!({})).await.unwrap();
+        assert_eq!(from_a.len(), 1);
+        assert_eq!(from_a[0].id, id_a);
+
+        let from_b = s.find(&cx_b, "shared", json!({})).await.unwrap();
+        assert_eq!(from_b.len(), 1);
+        assert_eq!(from_b[0].id, id_b);
+    }
+
+    #[tokio::test]
+    async fn anonymous_cx_skips_authz() {
+        // Denying authz repo + anon cx (no principal) ⇒ writes still
+        // succeed under script-as-gate.
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let id = s.create(&cx, "users", json!({ "x": 1 })).await.unwrap();
+        let _ = s.delete(&cx, "users", id).await.unwrap();
+    }
+
+    #[tokio::test]
+    async fn authed_cx_with_no_role_is_forbidden_on_write() {
+        let s = svc();
+        let cx = member_no_role_cx(AppId::new());
+        let err = s.create(&cx, "users", json!({ "x": 1 })).await.unwrap_err();
+        assert!(matches!(err, DocsError::Forbidden));
+    }
+
+    #[tokio::test]
+    async fn authed_cx_with_no_role_is_forbidden_on_read() {
+        let s = svc();
+        let cx = member_no_role_cx(AppId::new());
+        let err = s.get(&cx, "users", Uuid::new_v4()).await.unwrap_err();
+        assert!(matches!(err, DocsError::Forbidden));
+    }
+
+    #[tokio::test]
+    async fn owner_principal_can_write() {
+        let s = svc();
+        let cx = owner_cx(AppId::new());
+        let _ = s.create(&cx, "users", json!({ "x": 1 })).await.unwrap();
+    }
+
+    #[tokio::test]
+    async fn editor_member_can_write_via_role() {
+        // AllowingAuthzRepo grants Editor — should be able to write
+        // (AppDocsWrite is in_editor in role_satisfies).
+        let s = svc_allowing();
+        let cx = member_no_role_cx(AppId::new());
+        let _ = s.create(&cx, "users", json!({ "x": 1 })).await.unwrap();
+    }
+
+    #[tokio::test]
+    async fn find_with_equality_returns_matches() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        s.create(&cx, "users", json!({ "tier": "gold" }))
+            .await
+            .unwrap();
+        s.create(&cx, "users", json!({ "tier": "silver" }))
+            .await
+            .unwrap();
+        s.create(&cx, "users", json!({ "tier": "gold" }))
+            .await
+            .unwrap();
+
+        let golds = s
+            .find(&cx, "users", json!({ "tier": "gold" }))
+            .await
+            .unwrap();
+        assert_eq!(golds.len(), 2);
+    }
+
+    #[tokio::test]
+    async fn find_one_returns_first_or_none() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        s.create(&cx, "users", json!({ "tier": "gold" }))
+            .await
+            .unwrap();
+
+        let hit = s
+            .find_one(&cx, "users", json!({ "tier": "gold" }))
+            .await
+            .unwrap();
+        assert!(hit.is_some());
+
+        let miss = s
+            .find_one(&cx, "users", json!({ "tier": "platinum" }))
+            .await
+            .unwrap();
+        assert!(miss.is_none());
+    }
+
+    #[tokio::test]
+    async fn find_with_unsupported_operator_throws() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let err = s
+            .find(&cx, "users", json!({ "name": { "$regex": "^A" } }))
+            .await
+            .unwrap_err();
+        match err {
+            DocsError::UnsupportedOperator(m) => {
+                assert!(m.contains("$regex"));
+                assert!(m.contains("v1.2"));
+            }
+            other => panic!("expected UnsupportedOperator, got {other:?}"),
+        }
+    }
+
+    #[tokio::test]
+    async fn find_with_invalid_filter_throws() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let err = s
+            .find(&cx, "users", json!({ "a.b.c.d.e.f": "x" }))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, DocsError::InvalidFilter(_)));
+    }
+
+    #[tokio::test]
+    async fn find_with_dollar_in_returns_subset() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        s.create(&cx, "users", json!({ "tier": "gold" }))
+            .await
+            .unwrap();
+        s.create(&cx, "users", json!({ "tier": "silver" }))
+            .await
+            .unwrap();
+        s.create(&cx, "users", json!({ "tier": "platinum" }))
+            .await
+            .unwrap();
+
+        let hits = s
+            .find(
+                &cx,
+                "users",
+                json!({ "tier": { "$in": ["gold", "platinum"] } }),
+            )
+            .await
+            .unwrap();
+        assert_eq!(hits.len(), 2);
+    }
+
+    #[tokio::test]
+    async fn find_one_explicit_limit_is_honoured() {
+        // The service injects limit=1 ONLY when caller didn't set
+        // $limit. An explicit `$limit: 5` survives — and find_one
+        // still returns the first.
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        for _ in 0..3 {
+            s.create(&cx, "users", json!({ "tier": "gold" }))
+                .await
+                .unwrap();
+        }
+        let hit = s
+            .find_one(&cx, "users", json!({ "tier": "gold", "$limit": 5 }))
+            .await
+            .unwrap();
+        assert!(hit.is_some());
+    }
+
+    #[tokio::test]
+    async fn list_cursor_pagination() {
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let mut ids = Vec::new();
+        for _ in 0..5 {
+            ids.push(s.create(&cx, "users", json!({})).await.unwrap());
+        }
+        ids.sort();
+
+        let p1 = s.list(&cx, "users", None, 2).await.unwrap();
+        assert_eq!(p1.docs.len(), 2);
+        assert!(p1.next_cursor.is_some());
+
+        let p2 = s
+            .list(&cx, "users", p1.next_cursor.as_deref(), 2)
+            .await
+            .unwrap();
+        assert_eq!(p2.docs.len(), 2);
+
+        let p3 = s
+            .list(&cx, "users", p2.next_cursor.as_deref(), 2)
+            .await
+            .unwrap();
+        assert_eq!(p3.docs.len(), 1);
+        assert!(p3.next_cursor.is_none());
+    }
+
+    #[tokio::test]
+    async fn noop_emitter_does_not_block_mutations() {
+        // Pins v1.1.0 contract: services hold an Arc<dyn ServiceEventEmitter>
+        // and call emit().await unconditionally. The noop drops it.
+        let s = svc();
+        let cx = anon_cx(AppId::new());
+        let id = s.create(&cx, "users", json!({ "x": 1 })).await.unwrap();
+        s.update(&cx, "users", id, json!({ "x": 2 })).await.unwrap();
+        let _ = s.delete(&cx, "users", id).await.unwrap();
+    }
+}

crates/manager-core/src/files_api.rs

@@ -0,0 +1,215 @@
+//! `/api/v1/admin/apps/{id}/files*` — minimal files admin endpoints
+//! backing the dashboard's files view (v1.1.5).
+//!
+//! Two operations only, both operator-facing:
+//!   * `GET  /apps/{id}/files?collection=<c>&cursor=&limit=` — list file
+//!     metadata for a collection (cursor-paginated).
+//!   * `DELETE /apps/{id}/files/{collection}/{file_id}` — remove a file.
+//!
+//! These talk to the `FilesRepo` directly (like `triggers_api` talks to
+//! `TriggerRepo`), guarded by the same capability model as the SDK
+//! (`AppFilesRead` / `AppFilesWrite`). **Admin deletes do NOT emit a
+//! `files:delete` trigger event** — they're operator cleanup actions,
+//! not script mutations (see HANDBACK §7). The capability binds to the
+//! resource's `app_id` after the app is loaded.
+
+use std::sync::Arc;
+
+use axum::extract::{Path, Query, State};
+use axum::http::StatusCode;
+use axum::response::{IntoResponse, Json, Response};
+use axum::routing::{delete, get};
+use axum::{Extension, Router};
+use picloud_shared::{AppId, Principal};
+use serde::{Deserialize, Serialize};
+use serde_json::json;
+use uuid::Uuid;
+
+use crate::app_repo::AppRepository;
+use crate::authz::{require, AuthzDenied, AuthzError, AuthzRepo, Capability};
+use crate::files_repo::{FilesRepo, FilesRepoError};
+
+#[derive(Clone)]
+pub struct FilesAdminState {
+    pub files: Arc<dyn FilesRepo>,
+    pub apps: Arc<dyn AppRepository>,
+    pub authz: Arc<dyn AuthzRepo>,
+}
+
+pub fn files_admin_router(state: FilesAdminState) -> Router {
+    Router::new()
+        .route("/apps/{app_id}/files", get(list_files))
+        .route(
+            "/apps/{app_id}/files/{collection}/{file_id}",
+            delete(delete_file),
+        )
+        .with_state(state)
+}
+
+#[derive(Debug, Deserialize)]
+pub struct ListFilesQuery {
+    pub collection: String,
+    #[serde(default)]
+    pub cursor: Option<String>,
+    #[serde(default)]
+    pub limit: Option<u32>,
+}
+
+#[derive(Debug, Serialize)]
+struct FileMetaDto {
+    id: String,
+    collection: String,
+    name: String,
+    content_type: String,
+    size: u64,
+    checksum: String,
+    created_at: String,
+    updated_at: String,
+}
+
+#[derive(Debug, Serialize)]
+struct ListFilesResponse {
+    files: Vec<FileMetaDto>,
+    next_cursor: Option<String>,
+}
+
+async fn list_files(
+    State(s): State<FilesAdminState>,
+    Extension(principal): Extension<Principal>,
+    Path(app_id): Path<AppId>,
+    Query(q): Query<ListFilesQuery>,
+) -> Result<Json<ListFilesResponse>, FilesApiError> {
+    ensure_app_exists(&*s.apps, app_id).await?;
+    require(
+        s.authz.as_ref(),
+        &principal,
+        Capability::AppFilesRead(app_id),
+    )
+    .await?;
+    if q.collection.trim().is_empty() {
+        return Err(FilesApiError::Invalid(
+            "collection must not be empty".into(),
+        ));
+    }
+    let page = s
+        .files
+        .list(
+            app_id,
+            &q.collection,
+            q.cursor.as_deref(),
+            q.limit.unwrap_or(0),
+        )
+        .await?;
+    let files = page
+        .files
+        .into_iter()
+        .map(|m| FileMetaDto {
+            id: m.id.to_string(),
+            collection: m.collection,
+            name: m.name,
+            content_type: m.content_type,
+            size: m.size,
+            checksum: m.checksum,
+            created_at: m.created_at.to_rfc3339(),
+            updated_at: m.updated_at.to_rfc3339(),
+        })
+        .collect();
+    Ok(Json(ListFilesResponse {
+        files,
+        next_cursor: page.next_cursor,
+    }))
+}
+
+async fn delete_file(
+    State(s): State<FilesAdminState>,
+    Extension(principal): Extension<Principal>,
+    Path((app_id, collection, file_id)): Path<(AppId, String, String)>,
+) -> Result<StatusCode, FilesApiError> {
+    ensure_app_exists(&*s.apps, app_id).await?;
+    require(
+        s.authz.as_ref(),
+        &principal,
+        Capability::AppFilesWrite(app_id),
+    )
+    .await?;
+    let id = Uuid::parse_str(&file_id).map_err(|_| FilesApiError::NotFound)?;
+    if s.files.delete(app_id, &collection, id).await?.is_none() {
+        return Err(FilesApiError::NotFound);
+    }
+    Ok(StatusCode::NO_CONTENT)
+}
+
+async fn ensure_app_exists(apps: &dyn AppRepository, app_id: AppId) -> Result<(), FilesApiError> {
+    apps.get_by_id(app_id)
+        .await
+        .map_err(|e| FilesApiError::Backend(e.to_string()))?
+        .ok_or(FilesApiError::AppNotFound)?;
+    Ok(())
+}
+
+#[derive(Debug, thiserror::Error)]
+pub enum FilesApiError {
+    #[error("app not found")]
+    AppNotFound,
+    #[error("file not found")]
+    NotFound,
+    #[error("invalid request: {0}")]
+    Invalid(String),
+    #[error("forbidden")]
+    Forbidden,
+    #[error("authorization repo error: {0}")]
+    AuthzRepo(String),
+    #[error("files backend: {0}")]
+    Backend(String),
+}
+
+impl From<AuthzDenied> for FilesApiError {
+    fn from(d: AuthzDenied) -> Self {
+        match d {
+            AuthzDenied::Denied => Self::Forbidden,
+            AuthzDenied::Repo(e) => Self::AuthzRepo(e.to_string()),
+        }
+    }
+}
+
+impl From<AuthzError> for FilesApiError {
+    fn from(e: AuthzError) -> Self {
+        Self::AuthzRepo(e.to_string())
+    }
+}
+
+impl From<FilesRepoError> for FilesApiError {
+    fn from(e: FilesRepoError) -> Self {
+        Self::Backend(e.to_string())
+    }
+}
+
+impl IntoResponse for FilesApiError {
+    fn into_response(self) -> Response {
+        let (status, body) = match &self {
+            Self::AppNotFound | Self::NotFound => {
+                (StatusCode::NOT_FOUND, json!({ "error": self.to_string() }))
+            }
+            Self::Invalid(_) => (
+                StatusCode::UNPROCESSABLE_ENTITY,
+                json!({ "error": self.to_string() }),
+            ),
+            Self::Forbidden => (StatusCode::FORBIDDEN, json!({ "error": self.to_string() })),
+            Self::AuthzRepo(e) => {
+                tracing::error!(error = %e, "files admin authz repo error");
+                (
+                    StatusCode::INTERNAL_SERVER_ERROR,
+                    json!({ "error": "internal error" }),
+                )
+            }
+            Self::Backend(e) => {
+                tracing::error!(error = %e, "files admin backend error");
+                (
+                    StatusCode::INTERNAL_SERVER_ERROR,
+                    json!({ "error": "internal error" }),
+                )
+            }
+        };
+        (status, Json(body)).into_response()
+    }
+}

crates/manager-core/src/files_repo.rs

@@ -0,0 +1,759 @@
+//! `FilesRepo` — the metadata row (Postgres) + blob bytes (filesystem)
+//! storage layer for the v1.1.5 `files::*` SDK.
+//!
+//! Unlike KV/docs, this repo owns BOTH halves of a file: the `files`
+//! row (metadata + SHA-256 checksum) and the bytes on disk at
+//!   `<root>/files/<app_id>/<collection>/<id[0:2]>/<id>`.
+//! It owns both because the write must be atomic across them — a crash
+//! mid-write must never leave a readable half-written file.
+//!
+//! ## Atomic write protocol (`create` / `update`)
+//!   1. Validate (collection path-safety; caps live one layer up).
+//!   2. `create_dir_all` the shard dir with `0o700`.
+//!   3. SHA-256 the in-memory bytes (single pass) while writing to
+//!      `<final>.tmp.<unique>`.
+//!   4. `fsync` the temp file.
+//!   5. `rename` temp → final (atomic on POSIX).
+//!   6. `fsync` the parent dir (so the rename is durable).
+//!   7. INSERT / UPDATE the DB row.
+//!
+//! A crash between 1–5 leaves an orphan `*.tmp.*` (never read). A crash
+//! between 5–7 leaves a file with no row — never reachable via the SDK
+//! (reads start from the row). Both are reclaimed by a future orphan
+//! sweep (deferred to v1.1.6+; see HANDBACK §7).
+//!
+//! ## Atomic delete protocol
+//!   1. SELECT + DELETE the row inside one transaction; commit.
+//!   2. `unlink` the file (outside the tx). A failure here leaves an
+//!      orphan; a failure before the commit changes nothing.
+//!
+//! ## Checksum-on-read
+//! `get` reads the file, hashes it, and compares against the stored
+//! checksum — returning `FilesError::Corrupted` (and logging the path
+//! at error level) on a mismatch. It never auto-deletes; the operator
+//! decides what to do with a metadata-vs-bytes divergence.
+
+use std::env;
+use std::path::{Path, PathBuf};
+use std::sync::atomic::{AtomicU64, Ordering};
+
+use async_trait::async_trait;
+use base64::engine::general_purpose::URL_SAFE_NO_PAD;
+use base64::Engine as _;
+use chrono::{DateTime, Utc};
+use picloud_shared::{AppId, FileMeta, FileUpdate, FilesListPage, NewFile};
+use sha2::{Digest, Sha256};
+use sqlx::PgPool;
+use uuid::Uuid;
+
+/// 100 MB default per-file cap.
+pub const DEFAULT_MAX_FILE_SIZE_BYTES: usize = 100 * 1024 * 1024;
+/// Default filesystem root (relative to the process CWD).
+pub const DEFAULT_FILES_ROOT: &str = "./data";
+
+const FILES_LIST_MAX_LIMIT: u32 = 1_000;
+const FILES_LIST_DEFAULT_LIMIT: u32 = 100;
+
+/// Monotonic counter feeding unique temp-file suffixes (combined with
+/// the pid). Avoids `rand` in the storage layer per the brief.
+static TMP_COUNTER: AtomicU64 = AtomicU64::new(0);
+
+#[derive(Debug, thiserror::Error)]
+pub enum FilesRepoError {
+    #[error("database error: {0}")]
+    Db(#[from] sqlx::Error),
+
+    #[error("filesystem error: {0}")]
+    Io(String),
+
+    #[error("invalid collection name: {0}")]
+    InvalidCollection(String),
+
+    /// The bytes on disk no longer match the stored checksum (or are
+    /// missing entirely while the row persists).
+    #[error("file content corrupted (checksum mismatch)")]
+    Corrupted,
+
+    #[error("invalid pagination cursor")]
+    InvalidCursor,
+}
+
+/// Outbound-files tunables. Env-overridable following the same pattern
+/// as `HttpConfig::from_env`.
+#[derive(Debug, Clone)]
+pub struct FilesConfig {
+    pub root: PathBuf,
+    pub max_file_size_bytes: usize,
+}
+
+impl FilesConfig {
+    #[must_use]
+    pub fn conservative() -> Self {
+        Self {
+            root: PathBuf::from(DEFAULT_FILES_ROOT),
+            max_file_size_bytes: DEFAULT_MAX_FILE_SIZE_BYTES,
+        }
+    }
+
+    #[must_use]
+    pub fn from_env() -> Self {
+        let mut c = Self::conservative();
+        if let Ok(v) = env::var("PICLOUD_FILES_ROOT") {
+            if !v.trim().is_empty() {
+                c.root = PathBuf::from(v);
+            }
+        }
+        if let Ok(v) = env::var("PICLOUD_FILES_MAX_FILE_SIZE_BYTES") {
+            match v.parse::<usize>() {
+                Ok(n) => c.max_file_size_bytes = n,
+                Err(e) => {
+                    tracing::warn!(error = %e, "ignoring invalid PICLOUD_FILES_MAX_FILE_SIZE_BYTES");
+                }
+            }
+        }
+        c
+    }
+}
+
+impl Default for FilesConfig {
+    fn default() -> Self {
+        Self::conservative()
+    }
+}
+
+/// The new+prior metadata returned from a successful `update`, so the
+/// service can emit a `ServiceEvent` with the change-data-capture
+/// surface (`old_payload`).
+#[derive(Debug, Clone)]
+pub struct FileUpdated {
+    pub new: FileMeta,
+    pub prev: FileMeta,
+}
+
+#[async_trait]
+pub trait FilesRepo: Send + Sync {
+    async fn create(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        new: NewFile,
+    ) -> Result<FileMeta, FilesRepoError>;
+
+    async fn head(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: Uuid,
+    ) -> Result<Option<FileMeta>, FilesRepoError>;
+
+    /// Reads + checksum-verifies the bytes. `Ok(None)` when no row
+    /// exists; `Err(Corrupted)` when the row exists but the bytes are
+    /// missing or mismatched.
+    async fn get(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: Uuid,
+    ) -> Result<Option<Vec<u8>>, FilesRepoError>;
+
+    /// `Ok(None)` when no row exists (the SDK turns this into
+    /// `FilesError::NotFound`).
+    async fn update(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: Uuid,
+        upd: FileUpdate,
+    ) -> Result<Option<FileUpdated>, FilesRepoError>;
+
+    /// Returns the deleted row's metadata if present, `None` otherwise.
+    async fn delete(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: Uuid,
+    ) -> Result<Option<FileMeta>, FilesRepoError>;
+
+    async fn list(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        cursor: Option<&str>,
+        limit: u32,
+    ) -> Result<FilesListPage, FilesRepoError>;
+}
+
+/// Filesystem-bytes + Postgres-metadata repo.
+pub struct FsFilesRepo {
+    pool: PgPool,
+    config: FilesConfig,
+}
+
+impl FsFilesRepo {
+    #[must_use]
+    pub fn new(pool: PgPool, config: FilesConfig) -> Self {
+        Self { pool, config }
+    }
+
+    /// Defensive path-component guard. The service already validates the
+    /// collection at the SDK boundary; this is belt-and-suspenders so a
+    /// future caller can't smuggle a traversal sequence onto disk.
+    fn guard_collection(collection: &str) -> Result<(), FilesRepoError> {
+        if collection.is_empty()
+            || collection.contains('/')
+            || collection.contains('\\')
+            || collection.contains("..")
+            || collection.contains('\0')
+        {
+            return Err(FilesRepoError::InvalidCollection(collection.to_string()));
+        }
+        Ok(())
+    }
+
+    fn final_path(&self, app_id: AppId, collection: &str, id: Uuid) -> PathBuf {
+        final_path_at(&self.config.root, app_id, collection, id)
+    }
+
+    fn write_atomic(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: Uuid,
+        bytes: &[u8],
+    ) -> Result<String, FilesRepoError> {
+        write_atomic_at(&self.config.root, app_id, collection, id, bytes)
+    }
+}
+
+fn shard_dir_at(root: &Path, app_id: AppId, collection: &str, id_str: &str) -> PathBuf {
+    root.join("files")
+        .join(app_id.into_inner().to_string())
+        .join(collection)
+        .join(&id_str[..2])
+}
+
+fn final_path_at(root: &Path, app_id: AppId, collection: &str, id: Uuid) -> PathBuf {
+    let id_str = id.to_string();
+    shard_dir_at(root, app_id, collection, &id_str).join(&id_str)
+}
+
+/// Steps 2–6 of the atomic-write protocol. Returns the lowercase hex
+/// SHA-256 of the bytes (computed in a single pass over the in-memory
+/// buffer — the file is never re-read). Free function so the fs
+/// mechanics are unit-testable without a Postgres pool.
+fn write_atomic_at(
+    root: &Path,
+    app_id: AppId,
+    collection: &str,
+    id: Uuid,
+    bytes: &[u8],
+) -> Result<String, FilesRepoError> {
+    use std::io::Write as _;
+
+    let id_str = id.to_string();
+    let dir = shard_dir_at(root, app_id, collection, &id_str);
+    create_dir_all_secure(&dir)?;
+
+    // Single-pass checksum over the in-memory buffer.
+    let mut hasher = Sha256::new();
+    hasher.update(bytes);
+    let checksum = hex_lower(&hasher.finalize());
+
+    let seq = TMP_COUNTER.fetch_add(1, Ordering::Relaxed);
+    let tmp = dir.join(format!("{id_str}.tmp.{}-{seq}", std::process::id()));
+    let final_path = dir.join(&id_str);
+
+    {
+        let mut f = std::fs::File::create(&tmp).map_err(io_err)?;
+        f.write_all(bytes).map_err(io_err)?;
+        f.sync_all().map_err(io_err)?; // fsync temp
+    }
+    std::fs::rename(&tmp, &final_path).map_err(io_err)?; // atomic
+                                                         // fsync the parent dir so the rename is durable.
+    if let Ok(dirf) = std::fs::File::open(&dir) {
+        let _ = dirf.sync_all();
+    }
+    Ok(checksum)
+}
+
+/// Read + checksum-verify the bytes at the given path-set. Free
+/// function mirror of the `get` read path. Returns `Corrupted` when the
+/// bytes are missing or don't match `expected_checksum`.
+fn read_verify_at(
+    root: &Path,
+    app_id: AppId,
+    collection: &str,
+    id: Uuid,
+    expected_checksum: &str,
+) -> Result<Vec<u8>, FilesRepoError> {
+    let path = final_path_at(root, app_id, collection, id);
+    let bytes = match std::fs::read(&path) {
+        Ok(b) => b,
+        Err(e) => {
+            tracing::error!(
+                path = %path.display(), error = %e,
+                "files: row exists but bytes are unreadable — treating as corrupted"
+            );
+            return Err(FilesRepoError::Corrupted);
+        }
+    };
+    let mut hasher = Sha256::new();
+    hasher.update(&bytes);
+    let actual = hex_lower(&hasher.finalize());
+    if actual != expected_checksum {
+        tracing::error!(
+            path = %path.display(), expected = %expected_checksum, actual = %actual,
+            "files: checksum mismatch on read — content corrupted"
+        );
+        return Err(FilesRepoError::Corrupted);
+    }
+    Ok(bytes)
+}
+
+#[async_trait]
+impl FilesRepo for FsFilesRepo {
+    async fn create(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        new: NewFile,
+    ) -> Result<FileMeta, FilesRepoError> {
+        Self::guard_collection(collection)?;
+        let id = Uuid::new_v4();
+        let size = i64::try_from(new.data.len()).unwrap_or(i64::MAX);
+
+        let checksum = self.write_atomic(app_id, collection, id, &new.data)?;
+
+        let row: FileRow = sqlx::query_as(
+            "INSERT INTO files \
+                (app_id, collection, id, name, content_type, size_bytes, checksum_sha256) \
+             VALUES ($1, $2, $3, $4, $5, $6, $7) \
+             RETURNING id, collection, name, content_type, size_bytes, \
+                       checksum_sha256, created_at, updated_at",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(id)
+        .bind(&new.name)
+        .bind(&new.content_type)
+        .bind(size)
+        .bind(&checksum)
+        .fetch_one(&self.pool)
+        .await?;
+
+        Ok(row.into_meta())
+    }
+
+    async fn head(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: Uuid,
+    ) -> Result<Option<FileMeta>, FilesRepoError> {
+        let row: Option<FileRow> = sqlx::query_as(
+            "SELECT id, collection, name, content_type, size_bytes, \
+                    checksum_sha256, created_at, updated_at \
+             FROM files WHERE app_id = $1 AND collection = $2 AND id = $3",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(id)
+        .fetch_optional(&self.pool)
+        .await?;
+        Ok(row.map(FileRow::into_meta))
+    }
+
+    async fn get(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: Uuid,
+    ) -> Result<Option<Vec<u8>>, FilesRepoError> {
+        let row: Option<(String,)> = sqlx::query_as(
+            "SELECT checksum_sha256 FROM files \
+             WHERE app_id = $1 AND collection = $2 AND id = $3",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(id)
+        .fetch_optional(&self.pool)
+        .await?;
+        let Some((stored_checksum,)) = row else {
+            return Ok(None);
+        };
+        let bytes = read_verify_at(&self.config.root, app_id, collection, id, &stored_checksum)?;
+        Ok(Some(bytes))
+    }
+
+    async fn update(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: Uuid,
+        upd: FileUpdate,
+    ) -> Result<Option<FileUpdated>, FilesRepoError> {
+        Self::guard_collection(collection)?;
+        // Read the prior row first (existence check + CDC surface).
+        let Some(prev) = self.head(app_id, collection, id).await? else {
+            return Ok(None);
+        };
+
+        let size = i64::try_from(upd.data.len()).unwrap_or(i64::MAX);
+        let checksum = self.write_atomic(app_id, collection, id, &upd.data)?;
+
+        let row: FileRow = sqlx::query_as(
+            "UPDATE files SET \
+                name = COALESCE($4, name), \
+                content_type = COALESCE($5, content_type), \
+                size_bytes = $6, \
+                checksum_sha256 = $7, \
+                updated_at = NOW() \
+             WHERE app_id = $1 AND collection = $2 AND id = $3 \
+             RETURNING id, collection, name, content_type, size_bytes, \
+                       checksum_sha256, created_at, updated_at",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(id)
+        .bind(upd.name.as_deref())
+        .bind(upd.content_type.as_deref())
+        .bind(size)
+        .bind(&checksum)
+        .fetch_one(&self.pool)
+        .await?;
+
+        Ok(Some(FileUpdated {
+            new: row.into_meta(),
+            prev,
+        }))
+    }
+
+    async fn delete(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        id: Uuid,
+    ) -> Result<Option<FileMeta>, FilesRepoError> {
+        // SELECT + DELETE in one tx; unlink afterwards (outside the tx).
+        let mut tx = self.pool.begin().await?;
+        let row: Option<FileRow> = sqlx::query_as(
+            "SELECT id, collection, name, content_type, size_bytes, \
+                    checksum_sha256, created_at, updated_at \
+             FROM files WHERE app_id = $1 AND collection = $2 AND id = $3 \
+             FOR UPDATE",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(id)
+        .fetch_optional(&mut *tx)
+        .await?;
+
+        let Some(row) = row else {
+            tx.rollback().await?;
+            return Ok(None);
+        };
+
+        sqlx::query("DELETE FROM files WHERE app_id = $1 AND collection = $2 AND id = $3")
+            .bind(app_id.into_inner())
+            .bind(collection)
+            .bind(id)
+            .execute(&mut *tx)
+            .await?;
+        tx.commit().await?;
+
+        // Row is gone; unlink the bytes. A failure here leaves an orphan
+        // file (reclaimed by a future sweep) — not fatal.
+        let path = self.final_path(app_id, collection, id);
+        if let Err(e) = std::fs::remove_file(&path) {
+            if e.kind() != std::io::ErrorKind::NotFound {
+                tracing::warn!(path = %path.display(), error = %e, "files: unlink after delete failed (orphan)");
+            }
+        }
+        Ok(Some(row.into_meta()))
+    }
+
+    async fn list(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        cursor: Option<&str>,
+        limit: u32,
+    ) -> Result<FilesListPage, FilesRepoError> {
+        let limit = if limit == 0 {
+            FILES_LIST_DEFAULT_LIMIT
+        } else {
+            limit.min(FILES_LIST_MAX_LIMIT)
+        };
+        let last_id = match cursor {
+            Some(c) => Some(decode_cursor(c)?),
+            None => None,
+        };
+        let take = i64::from(limit) + 1;
+        let rows: Vec<FileRow> = sqlx::query_as(
+            "SELECT id, collection, name, content_type, size_bytes, \
+                    checksum_sha256, created_at, updated_at \
+             FROM files \
+             WHERE app_id = $1 AND collection = $2 \
+               AND ($3::uuid IS NULL OR id > $3) \
+             ORDER BY id ASC \
+             LIMIT $4",
+        )
+        .bind(app_id.into_inner())
+        .bind(collection)
+        .bind(last_id)
+        .bind(take)
+        .fetch_all(&self.pool)
+        .await?;
+
+        let mut files: Vec<FileMeta> = rows.into_iter().map(FileRow::into_meta).collect();
+        let next_cursor = if files.len() > limit as usize {
+            files.truncate(limit as usize);
+            files.last().map(|m| encode_cursor(m.id))
+        } else {
+            None
+        };
+        Ok(FilesListPage { files, next_cursor })
+    }
+}
+
+// ----------------------------------------------------------------------------
+// Helpers
+// ----------------------------------------------------------------------------
+
+fn io_err(e: std::io::Error) -> FilesRepoError {
+    FilesRepoError::Io(e.to_string())
+}
+
+/// `create_dir_all` with `0o700` on the created tree (Unix). On other
+/// platforms it falls back to the default permissions.
+fn create_dir_all_secure(dir: &Path) -> Result<(), FilesRepoError> {
+    #[cfg(unix)]
+    {
+        use std::os::unix::fs::DirBuilderExt as _;
+        std::fs::DirBuilder::new()
+            .recursive(true)
+            .mode(0o700)
+            .create(dir)
+            .map_err(io_err)
+    }
+    #[cfg(not(unix))]
+    {
+        std::fs::create_dir_all(dir).map_err(io_err)
+    }
+}
+
+fn hex_lower(bytes: &[u8]) -> String {
+    let mut s = String::with_capacity(bytes.len() * 2);
+    for b in bytes {
+        use std::fmt::Write as _;
+        let _ = write!(s, "{b:02x}");
+    }
+    s
+}
+
+fn encode_cursor(last_id: Uuid) -> String {
+    URL_SAFE_NO_PAD.encode(last_id.to_string().as_bytes())
+}
+
+fn decode_cursor(cursor: &str) -> Result<Uuid, FilesRepoError> {
+    let bytes = URL_SAFE_NO_PAD
+        .decode(cursor)
+        .map_err(|_| FilesRepoError::InvalidCursor)?;
+    let s = String::from_utf8(bytes).map_err(|_| FilesRepoError::InvalidCursor)?;
+    Uuid::parse_str(&s).map_err(|_| FilesRepoError::InvalidCursor)
+}
+
+#[derive(sqlx::FromRow)]
+struct FileRow {
+    id: Uuid,
+    collection: String,
+    name: String,
+    content_type: String,
+    size_bytes: i64,
+    checksum_sha256: String,
+    created_at: DateTime<Utc>,
+    updated_at: DateTime<Utc>,
+}
+
+impl FileRow {
+    fn into_meta(self) -> FileMeta {
+        FileMeta {
+            id: self.id,
+            collection: self.collection,
+            name: self.name,
+            content_type: self.content_type,
+            size: u64::try_from(self.size_bytes).unwrap_or(0),
+            checksum: self.checksum_sha256,
+            created_at: self.created_at,
+            updated_at: self.updated_at,
+        }
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn hex_lower_matches_known_sha256_vector() {
+        // SHA-256("abc") — NIST known-answer vector.
+        let mut h = Sha256::new();
+        h.update(b"abc");
+        assert_eq!(
+            hex_lower(&h.finalize()),
+            "ba7816bf8f01cfea414140de5dae2223b00361a396177a9cb410ff61f20015ad"
+        );
+    }
+
+    #[test]
+    fn hex_lower_of_empty_is_known_vector() {
+        let mut h = Sha256::new();
+        h.update(b"");
+        assert_eq!(
+            hex_lower(&h.finalize()),
+            "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"
+        );
+    }
+
+    #[test]
+    fn cursor_round_trips() {
+        let id = Uuid::new_v4();
+        let enc = encode_cursor(id);
+        assert_eq!(decode_cursor(&enc).unwrap(), id);
+        assert!(matches!(
+            decode_cursor("!!not-base64!!"),
+            Err(FilesRepoError::InvalidCursor)
+        ));
+    }
+
+    #[test]
+    fn guard_collection_rejects_traversal() {
+        assert!(FsFilesRepo::guard_collection("avatars").is_ok());
+        assert!(FsFilesRepo::guard_collection("a/b").is_err());
+        assert!(FsFilesRepo::guard_collection("..").is_err());
+        assert!(FsFilesRepo::guard_collection("a..b").is_err());
+        assert!(FsFilesRepo::guard_collection("").is_err());
+        assert!(FsFilesRepo::guard_collection("a\0b").is_err());
+    }
+
+    #[test]
+    fn config_from_env_defaults_are_conservative() {
+        let c = FilesConfig::conservative();
+        assert_eq!(c.max_file_size_bytes, DEFAULT_MAX_FILE_SIZE_BYTES);
+        assert_eq!(c.root, PathBuf::from(DEFAULT_FILES_ROOT));
+    }
+
+    // ------------------------------------------------------------------
+    // Tempdir-backed filesystem mechanics — exercise the atomic write,
+    // single-pass checksum, and checksum-on-read tamper detection
+    // without needing a Postgres pool.
+    // ------------------------------------------------------------------
+
+    use picloud_shared::AppId;
+
+    /// Process-unique scratch dir under the system temp dir. Cleaned up
+    /// by each test via `remove_dir_all`.
+    fn unique_tmp_root() -> PathBuf {
+        let seq = TMP_COUNTER.fetch_add(1, Ordering::Relaxed);
+        let dir =
+            std::env::temp_dir().join(format!("picloud-files-test-{}-{seq}", std::process::id()));
+        std::fs::create_dir_all(&dir).unwrap();
+        dir
+    }
+
+    #[test]
+    fn write_atomic_then_read_verify_round_trips() {
+        let root = unique_tmp_root();
+        let app = AppId::new();
+        let id = Uuid::new_v4();
+        let bytes = b"hello picloud files".to_vec();
+
+        let checksum = write_atomic_at(&root, app, "avatars", id, &bytes).unwrap();
+        // Single-pass checksum matches an independent hash of the bytes.
+        let mut h = Sha256::new();
+        h.update(&bytes);
+        assert_eq!(checksum, hex_lower(&h.finalize()));
+
+        let read = read_verify_at(&root, app, "avatars", id, &checksum).unwrap();
+        assert_eq!(read, bytes);
+
+        std::fs::remove_dir_all(&root).ok();
+    }
+
+    #[test]
+    fn read_verify_detects_tampering_as_corrupted() {
+        let root = unique_tmp_root();
+        let app = AppId::new();
+        let id = Uuid::new_v4();
+        let checksum = write_atomic_at(&root, app, "c", id, b"original").unwrap();
+
+        // Mutate the bytes behind the repo's back.
+        let path = final_path_at(&root, app, "c", id);
+        std::fs::write(&path, b"tampered").unwrap();
+
+        let err = read_verify_at(&root, app, "c", id, &checksum).unwrap_err();
+        assert!(matches!(err, FilesRepoError::Corrupted));
+
+        std::fs::remove_dir_all(&root).ok();
+    }
+
+    #[test]
+    fn read_verify_missing_bytes_is_corrupted() {
+        let root = unique_tmp_root();
+        let app = AppId::new();
+        let id = Uuid::new_v4();
+        // No write — the file never existed.
+        let err = read_verify_at(&root, app, "c", id, "deadbeef").unwrap_err();
+        assert!(matches!(err, FilesRepoError::Corrupted));
+        std::fs::remove_dir_all(&root).ok();
+    }
+
+    #[test]
+    fn atomic_write_leaves_no_tmp_file_after_success() {
+        let root = unique_tmp_root();
+        let app = AppId::new();
+        let id = Uuid::new_v4();
+        write_atomic_at(&root, app, "c", id, b"data").unwrap();
+
+        let id_str = id.to_string();
+        let dir = shard_dir_at(&root, app, "c", &id_str);
+        let entries: Vec<_> = std::fs::read_dir(&dir)
+            .unwrap()
+            .filter_map(Result::ok)
+            .map(|e| e.file_name().to_string_lossy().into_owned())
+            .collect();
+        // Exactly the final file is visible — no `*.tmp.*` orphan.
+        assert_eq!(entries, vec![id_str]);
+        assert!(!entries.iter().any(|n| n.contains(".tmp.")));
+
+        std::fs::remove_dir_all(&root).ok();
+    }
+
+    #[test]
+    fn id_shard_uses_first_two_chars() {
+        let root = PathBuf::from("/tmp/x");
+        let app = AppId::new();
+        let id = Uuid::new_v4();
+        let id_str = id.to_string();
+        let path = final_path_at(&root, app, "col", id);
+        let shard = &id_str[..2];
+        assert!(path
+            .to_string_lossy()
+            .contains(&format!("/col/{shard}/{id_str}")));
+    }
+
+    #[cfg(unix)]
+    #[test]
+    fn shard_tree_created_with_0700() {
+        use std::os::unix::fs::PermissionsExt as _;
+        let root = unique_tmp_root();
+        let app = AppId::new();
+        let id = Uuid::new_v4();
+        write_atomic_at(&root, app, "c", id, b"data").unwrap();
+        let id_str = id.to_string();
+        let dir = shard_dir_at(&root, app, "c", &id_str);
+        let mode = std::fs::metadata(&dir).unwrap().permissions().mode();
+        assert_eq!(mode & 0o777, 0o700, "shard dir should be 0o700");
+        std::fs::remove_dir_all(&root).ok();
+    }
+}

crates/manager-core/src/files_service.rs

@@ -0,0 +1,817 @@
+//! `FilesServiceImpl` — wires the `FilesRepo` underneath the
+//! `picloud_shared::FilesService` trait scripts see via the Rhai
+//! bridge.
+//!
+//! Layers added here (vs the raw repo), mirroring `KvServiceImpl`:
+//!   1. Collection validation (empty + path-traversal) and field /
+//!      size-cap validation at the SDK boundary.
+//!   2. **Script-as-gate authz**: when `cx.principal.is_some()` we run
+//!      `authz::require(...)`; when it's `None` (public HTTP) we skip.
+//!      Cross-app isolation is unaffected — every repo call is keyed by
+//!      `cx.app_id`, never an argument.
+//!   3. `ServiceEvent` emission after each mutation (`create` /
+//!      `update` / `delete`). The payload is the file **metadata**, not
+//!      the blob bytes (files are too big for trigger payloads).
+
+use std::sync::Arc;
+
+use async_trait::async_trait;
+use picloud_shared::{
+    validate_files_collection, FileMeta, FileUpdate, FilesError, FilesListPage, FilesService,
+    NewFile, SdkCallCx, ServiceEvent, ServiceEventEmitter,
+};
+use uuid::Uuid;
+
+use crate::authz::{self, AuthzRepo, Capability};
+use crate::files_repo::{FileUpdated, FilesRepo, FilesRepoError};
+
+pub struct FilesServiceImpl {
+    repo: Arc<dyn FilesRepo>,
+    authz: Arc<dyn AuthzRepo>,
+    events: Arc<dyn ServiceEventEmitter>,
+    max_file_size_bytes: usize,
+}
+
+impl FilesServiceImpl {
+    #[must_use]
+    pub fn new(
+        repo: Arc<dyn FilesRepo>,
+        authz: Arc<dyn AuthzRepo>,
+        events: Arc<dyn ServiceEventEmitter>,
+        max_file_size_bytes: usize,
+    ) -> Self {
+        Self {
+            repo,
+            authz,
+            events,
+            max_file_size_bytes,
+        }
+    }
+
+    async fn check_read(&self, cx: &SdkCallCx) -> Result<(), FilesError> {
+        if let Some(ref principal) = cx.principal {
+            authz::require(&*self.authz, principal, Capability::AppFilesRead(cx.app_id))
+                .await
+                .map_err(|_| FilesError::Forbidden)?;
+        }
+        Ok(())
+    }
+
+    async fn check_write(&self, cx: &SdkCallCx) -> Result<(), FilesError> {
+        if let Some(ref principal) = cx.principal {
+            authz::require(
+                &*self.authz,
+                principal,
+                Capability::AppFilesWrite(cx.app_id),
+            )
+            .await
+            .map_err(|_| FilesError::Forbidden)?;
+        }
+        Ok(())
+    }
+
+    /// Best-effort `ServiceEvent` emission. A failed emit is logged but
+    /// never rolls back the (already-durable) file write.
+    async fn emit(
+        &self,
+        cx: &SdkCallCx,
+        op: &'static str,
+        collection: &str,
+        meta: &FileMeta,
+        old: Option<&FileMeta>,
+    ) {
+        let payload = serde_json::to_value(meta).ok();
+        let old_payload = old.and_then(|m| serde_json::to_value(m).ok());
+        if let Err(e) = self
+            .events
+            .emit(
+                cx,
+                ServiceEvent {
+                    source: "files",
+                    op,
+                    collection: Some(collection.to_string()),
+                    key: Some(meta.id.to_string()),
+                    payload,
+                    old_payload,
+                },
+            )
+            .await
+        {
+            tracing::warn!(error = %e, source = "files", op, "event emit failed");
+        }
+    }
+}
+
+/// Parse a script-supplied id. Invalid UUIDs aren't an error shape the
+/// SDK exposes — for reads/deletes they simply mean "no such file".
+fn parse_id(id: &str) -> Option<Uuid> {
+    Uuid::parse_str(id).ok()
+}
+
+impl From<FilesRepoError> for FilesError {
+    fn from(e: FilesRepoError) -> Self {
+        match e {
+            FilesRepoError::Corrupted => Self::Corrupted,
+            FilesRepoError::InvalidCollection(c) => Self::InvalidCollection(c),
+            other => Self::Backend(other.to_string()),
+        }
+    }
+}
+
+#[async_trait]
+impl FilesService for FilesServiceImpl {
+    async fn create(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        new: NewFile,
+    ) -> Result<Uuid, FilesError> {
+        validate_files_collection(collection)?;
+        self.check_write(cx).await?;
+        new.validate(self.max_file_size_bytes)?;
+        let meta = self.repo.create(cx.app_id, collection, new).await?;
+        self.emit(cx, "create", collection, &meta, None).await;
+        Ok(meta.id)
+    }
+
+    async fn head(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: &str,
+    ) -> Result<Option<FileMeta>, FilesError> {
+        validate_files_collection(collection)?;
+        self.check_read(cx).await?;
+        let Some(uuid) = parse_id(id) else {
+            return Ok(None);
+        };
+        Ok(self.repo.head(cx.app_id, collection, uuid).await?)
+    }
+
+    async fn get(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: &str,
+    ) -> Result<Option<Vec<u8>>, FilesError> {
+        validate_files_collection(collection)?;
+        self.check_read(cx).await?;
+        let Some(uuid) = parse_id(id) else {
+            return Ok(None);
+        };
+        Ok(self.repo.get(cx.app_id, collection, uuid).await?)
+    }
+
+    async fn update(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: &str,
+        upd: FileUpdate,
+    ) -> Result<(), FilesError> {
+        validate_files_collection(collection)?;
+        self.check_write(cx).await?;
+        upd.validate(self.max_file_size_bytes)?;
+        let Some(uuid) = parse_id(id) else {
+            return Err(FilesError::NotFound);
+        };
+        match self.repo.update(cx.app_id, collection, uuid, upd).await? {
+            Some(FileUpdated { new, prev }) => {
+                self.emit(cx, "update", collection, &new, Some(&prev)).await;
+                Ok(())
+            }
+            None => Err(FilesError::NotFound),
+        }
+    }
+
+    async fn delete(&self, cx: &SdkCallCx, collection: &str, id: &str) -> Result<bool, FilesError> {
+        validate_files_collection(collection)?;
+        self.check_write(cx).await?;
+        let Some(uuid) = parse_id(id) else {
+            return Ok(false);
+        };
+        match self.repo.delete(cx.app_id, collection, uuid).await? {
+            Some(meta) => {
+                // On delete, the top-level metadata AND `prev` both carry
+                // the deleted row (per docs/v1.1.x design + the brief).
+                self.emit(cx, "delete", collection, &meta, Some(&meta))
+                    .await;
+                Ok(true)
+            }
+            None => Ok(false),
+        }
+    }
+
+    async fn list(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        cursor: Option<&str>,
+        limit: u32,
+    ) -> Result<FilesListPage, FilesError> {
+        validate_files_collection(collection)?;
+        self.check_read(cx).await?;
+        Ok(self.repo.list(cx.app_id, collection, cursor, limit).await?)
+    }
+}
+
+// ----------------------------------------------------------------------------
+// Tests — in-memory FilesRepo so unit tests need neither Postgres nor a
+// filesystem. The on-disk atomic-write / checksum mechanics are covered
+// by the tempdir tests in `files_repo.rs`.
+// ----------------------------------------------------------------------------
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::authz::{AuthzError, AuthzRepo};
+    use async_trait::async_trait;
+    use chrono::Utc;
+    use picloud_shared::{
+        AdminUserId, AppId, AppRole, EmitError, ExecutionId, InstanceRole, Principal, RequestId,
+        ScriptId, ServiceEvent, UserId,
+    };
+    use std::collections::BTreeMap;
+    use std::sync::Mutex as StdMutex;
+    use tokio::sync::Mutex;
+
+    /// In-memory FilesRepo keyed by (app, collection, id). Stores the
+    /// metadata + bytes together so cross-app isolation and round-trips
+    /// can be checked without disk.
+    #[derive(Default)]
+    struct InMemoryFilesRepo {
+        #[allow(clippy::type_complexity)]
+        data: Mutex<BTreeMap<(AppId, String, Uuid), (FileMeta, Vec<u8>)>>,
+    }
+
+    fn sha256_hex(bytes: &[u8]) -> String {
+        use sha2::{Digest, Sha256};
+        let mut h = Sha256::new();
+        h.update(bytes);
+        let out = h.finalize();
+        let mut s = String::new();
+        for b in out {
+            use std::fmt::Write as _;
+            let _ = write!(s, "{b:02x}");
+        }
+        s
+    }
+
+    #[async_trait]
+    impl FilesRepo for InMemoryFilesRepo {
+        async fn create(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            new: NewFile,
+        ) -> Result<FileMeta, FilesRepoError> {
+            let id = Uuid::new_v4();
+            let now = Utc::now();
+            let meta = FileMeta {
+                id,
+                collection: collection.to_string(),
+                name: new.name.clone(),
+                content_type: new.content_type.clone(),
+                size: new.data.len() as u64,
+                checksum: sha256_hex(&new.data),
+                created_at: now,
+                updated_at: now,
+            };
+            self.data.lock().await.insert(
+                (app_id, collection.to_string(), id),
+                (meta.clone(), new.data),
+            );
+            Ok(meta)
+        }
+
+        async fn head(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            id: Uuid,
+        ) -> Result<Option<FileMeta>, FilesRepoError> {
+            Ok(self
+                .data
+                .lock()
+                .await
+                .get(&(app_id, collection.to_string(), id))
+                .map(|(m, _)| m.clone()))
+        }
+
+        async fn get(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            id: Uuid,
+        ) -> Result<Option<Vec<u8>>, FilesRepoError> {
+            Ok(self
+                .data
+                .lock()
+                .await
+                .get(&(app_id, collection.to_string(), id))
+                .map(|(_, b)| b.clone()))
+        }
+
+        async fn update(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            id: Uuid,
+            upd: FileUpdate,
+        ) -> Result<Option<FileUpdated>, FilesRepoError> {
+            let mut data = self.data.lock().await;
+            let key = (app_id, collection.to_string(), id);
+            let Some((prev_meta, _)) = data.get(&key).cloned() else {
+                return Ok(None);
+            };
+            let now = Utc::now();
+            let new_meta = FileMeta {
+                id,
+                collection: collection.to_string(),
+                name: upd.name.clone().unwrap_or_else(|| prev_meta.name.clone()),
+                content_type: upd
+                    .content_type
+                    .clone()
+                    .unwrap_or_else(|| prev_meta.content_type.clone()),
+                size: upd.data.len() as u64,
+                checksum: sha256_hex(&upd.data),
+                created_at: prev_meta.created_at,
+                updated_at: now,
+            };
+            data.insert(key, (new_meta.clone(), upd.data));
+            Ok(Some(FileUpdated {
+                new: new_meta,
+                prev: prev_meta,
+            }))
+        }
+
+        async fn delete(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            id: Uuid,
+        ) -> Result<Option<FileMeta>, FilesRepoError> {
+            Ok(self
+                .data
+                .lock()
+                .await
+                .remove(&(app_id, collection.to_string(), id))
+                .map(|(m, _)| m))
+        }
+
+        async fn list(
+            &self,
+            app_id: AppId,
+            collection: &str,
+            cursor: Option<&str>,
+            limit: u32,
+        ) -> Result<FilesListPage, FilesRepoError> {
+            let data = self.data.lock().await;
+            let after = cursor.and_then(|c| Uuid::parse_str(c).ok());
+            let mut metas: Vec<FileMeta> = data
+                .iter()
+                .filter(|((a, c, _), _)| *a == app_id && c == collection)
+                .map(|(_, (m, _))| m.clone())
+                .filter(|m| after.is_none_or(|a| m.id > a))
+                .collect();
+            metas.sort_by_key(|m| m.id);
+            let take = (limit.max(1)) as usize;
+            let next_cursor = if metas.len() > take {
+                metas.truncate(take);
+                metas.last().map(|m| m.id.to_string())
+            } else {
+                None
+            };
+            Ok(FilesListPage {
+                files: metas,
+                next_cursor,
+            })
+        }
+    }
+
+    /// Captures emitted events so tests can assert on fan-out shape.
+    #[derive(Default)]
+    struct CapturingEmitter {
+        events: StdMutex<Vec<ServiceEvent>>,
+    }
+
+    #[async_trait]
+    impl ServiceEventEmitter for CapturingEmitter {
+        async fn emit(&self, _cx: &SdkCallCx, event: ServiceEvent) -> Result<(), EmitError> {
+            self.events.lock().unwrap().push(event);
+            Ok(())
+        }
+    }
+
+    #[derive(Default)]
+    struct DenyingAuthzRepo;
+    #[async_trait]
+    impl AuthzRepo for DenyingAuthzRepo {
+        async fn membership(
+            &self,
+            _user_id: UserId,
+            _app_id: AppId,
+        ) -> Result<Option<AppRole>, AuthzError> {
+            Ok(None)
+        }
+    }
+
+    #[derive(Default)]
+    struct EditorAuthzRepo;
+    #[async_trait]
+    impl AuthzRepo for EditorAuthzRepo {
+        async fn membership(
+            &self,
+            _user_id: UserId,
+            _app_id: AppId,
+        ) -> Result<Option<AppRole>, AuthzError> {
+            Ok(Some(AppRole::Editor))
+        }
+    }
+
+    fn anon_cx(app_id: AppId) -> SdkCallCx {
+        SdkCallCx {
+            app_id,
+            script_id: ScriptId::new(),
+            principal: None,
+            execution_id: ExecutionId::new(),
+            request_id: RequestId::new(),
+            trigger_depth: 0,
+            root_execution_id: ExecutionId::new(),
+            is_dead_letter_handler: false,
+            event: None,
+        }
+    }
+
+    fn member_cx(app_id: AppId) -> SdkCallCx {
+        SdkCallCx {
+            principal: Some(Principal {
+                user_id: AdminUserId::new(),
+                instance_role: InstanceRole::Member,
+                scopes: None,
+                app_binding: None,
+            }),
+            ..anon_cx(app_id)
+        }
+    }
+
+    fn svc_with(authz: Arc<dyn AuthzRepo>, emitter: Arc<CapturingEmitter>) -> FilesServiceImpl {
+        FilesServiceImpl::new(
+            Arc::new(InMemoryFilesRepo::default()),
+            authz,
+            emitter,
+            10 * 1024 * 1024,
+        )
+    }
+
+    fn svc() -> FilesServiceImpl {
+        svc_with(
+            Arc::new(DenyingAuthzRepo),
+            Arc::new(CapturingEmitter::default()),
+        )
+    }
+
+    fn new_file(name: &str, data: &[u8]) -> NewFile {
+        NewFile {
+            name: name.to_string(),
+            content_type: "application/octet-stream".to_string(),
+            data: data.to_vec(),
+        }
+    }
+
+    #[tokio::test]
+    async fn create_then_get_head_round_trips() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        let id = files
+            .create(&cx, "avatars", new_file("a.bin", b"hello"))
+            .await
+            .unwrap();
+        let bytes = files.get(&cx, "avatars", &id.to_string()).await.unwrap();
+        assert_eq!(bytes, Some(b"hello".to_vec()));
+        let meta = files
+            .head(&cx, "avatars", &id.to_string())
+            .await
+            .unwrap()
+            .unwrap();
+        assert_eq!(meta.name, "a.bin");
+        assert_eq!(meta.size, 5);
+        assert_eq!(meta.checksum, sha256_hex(b"hello"));
+    }
+
+    #[tokio::test]
+    async fn get_and_head_missing_return_none() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        let missing = Uuid::new_v4().to_string();
+        assert_eq!(files.get(&cx, "c", &missing).await.unwrap(), None);
+        assert!(files.head(&cx, "c", &missing).await.unwrap().is_none());
+        // Non-UUID id is also "missing", not an error.
+        assert_eq!(files.get(&cx, "c", "not-a-uuid").await.unwrap(), None);
+    }
+
+    #[tokio::test]
+    async fn update_replaces_content_and_keeps_metadata_when_omitted() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        let id = files
+            .create(&cx, "c", new_file("v1.txt", b"one"))
+            .await
+            .unwrap();
+        files
+            .update(
+                &cx,
+                "c",
+                &id.to_string(),
+                FileUpdate {
+                    data: b"two!!".to_vec(),
+                    name: None,
+                    content_type: None,
+                },
+            )
+            .await
+            .unwrap();
+        let meta = files
+            .head(&cx, "c", &id.to_string())
+            .await
+            .unwrap()
+            .unwrap();
+        assert_eq!(meta.name, "v1.txt"); // kept
+        assert_eq!(meta.size, 5);
+        assert_eq!(
+            files.get(&cx, "c", &id.to_string()).await.unwrap(),
+            Some(b"two!!".to_vec())
+        );
+    }
+
+    #[tokio::test]
+    async fn update_missing_throws_not_found() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        let err = files
+            .update(
+                &cx,
+                "c",
+                &Uuid::new_v4().to_string(),
+                FileUpdate {
+                    data: b"x".to_vec(),
+                    name: None,
+                    content_type: None,
+                },
+            )
+            .await
+            .unwrap_err();
+        assert!(matches!(err, FilesError::NotFound));
+    }
+
+    #[tokio::test]
+    async fn delete_returns_was_present() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        let id = files.create(&cx, "c", new_file("f", b"x")).await.unwrap();
+        assert!(files.delete(&cx, "c", &id.to_string()).await.unwrap());
+        assert!(!files.delete(&cx, "c", &id.to_string()).await.unwrap());
+        assert!(!files.delete(&cx, "c", "not-a-uuid").await.unwrap());
+    }
+
+    #[tokio::test]
+    async fn empty_collection_rejected() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        let err = files
+            .create(&cx, "", new_file("f", b"x"))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, FilesError::InvalidCollection(_)));
+    }
+
+    #[tokio::test]
+    async fn traversal_collection_rejected() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        for bad in ["../etc", "a/b", "a..b", "x\0y"] {
+            let err = files
+                .create(&cx, bad, new_file("f", b"x"))
+                .await
+                .unwrap_err();
+            assert!(
+                matches!(err, FilesError::InvalidCollection(_)),
+                "expected reject for {bad:?}"
+            );
+        }
+    }
+
+    #[tokio::test]
+    async fn missing_required_fields_have_field_specific_messages() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        // name
+        let err = files
+            .create(
+                &cx,
+                "c",
+                NewFile {
+                    name: "  ".into(),
+                    content_type: "text/plain".into(),
+                    data: b"x".to_vec(),
+                },
+            )
+            .await
+            .unwrap_err();
+        assert!(matches!(err, FilesError::MissingField("name")));
+        // content_type
+        let err = files
+            .create(
+                &cx,
+                "c",
+                NewFile {
+                    name: "f".into(),
+                    content_type: String::new(),
+                    data: b"x".to_vec(),
+                },
+            )
+            .await
+            .unwrap_err();
+        assert!(matches!(err, FilesError::MissingField("content_type")));
+        // data
+        let err = files
+            .create(
+                &cx,
+                "c",
+                NewFile {
+                    name: "f".into(),
+                    content_type: "text/plain".into(),
+                    data: vec![],
+                },
+            )
+            .await
+            .unwrap_err();
+        assert!(matches!(err, FilesError::MissingField("data")));
+    }
+
+    #[tokio::test]
+    async fn name_and_content_type_length_caps_enforced() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        let long_name = "x".repeat(256);
+        let err = files
+            .create(&cx, "c", new_file(&long_name, b"x"))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, FilesError::NameTooLong(256)));
+
+        let err = files
+            .create(
+                &cx,
+                "c",
+                NewFile {
+                    name: "f".into(),
+                    content_type: "x".repeat(128),
+                    data: b"x".to_vec(),
+                },
+            )
+            .await
+            .unwrap_err();
+        assert!(matches!(err, FilesError::ContentTypeTooLong(128)));
+    }
+
+    #[tokio::test]
+    async fn per_file_size_cap_enforced() {
+        let files = FilesServiceImpl::new(
+            Arc::new(InMemoryFilesRepo::default()),
+            Arc::new(DenyingAuthzRepo),
+            Arc::new(CapturingEmitter::default()),
+            8, // tiny cap
+        );
+        let cx = anon_cx(AppId::new());
+        let err = files
+            .create(&cx, "c", new_file("big", b"123456789"))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, FilesError::TooLarge { limit: 8, .. }));
+    }
+
+    #[tokio::test]
+    async fn cross_app_isolation() {
+        let files = svc();
+        let app_a = AppId::new();
+        let app_b = AppId::new();
+        let cx_a = anon_cx(app_a);
+        let cx_b = anon_cx(app_b);
+        let id = files
+            .create(&cx_a, "shared", new_file("f", b"from-a"))
+            .await
+            .unwrap();
+        // app B cannot see app A's file by id.
+        assert_eq!(
+            files.get(&cx_b, "shared", &id.to_string()).await.unwrap(),
+            None
+        );
+        assert!(files
+            .head(&cx_b, "shared", &id.to_string())
+            .await
+            .unwrap()
+            .is_none());
+        let page_b = files.list(&cx_b, "shared", None, 100).await.unwrap();
+        assert!(page_b.files.is_empty());
+        // app A still sees it.
+        assert!(files
+            .get(&cx_a, "shared", &id.to_string())
+            .await
+            .unwrap()
+            .is_some());
+    }
+
+    #[tokio::test]
+    async fn anonymous_cx_skips_authz() {
+        let files = svc(); // DenyingAuthzRepo
+        let cx = anon_cx(AppId::new());
+        // No principal → no authz check, even with a denying repo.
+        files.create(&cx, "c", new_file("f", b"x")).await.unwrap();
+    }
+
+    #[tokio::test]
+    async fn member_without_role_is_forbidden() {
+        let files = svc(); // DenyingAuthzRepo
+        let cx = member_cx(AppId::new());
+        let err = files
+            .create(&cx, "c", new_file("f", b"x"))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, FilesError::Forbidden));
+    }
+
+    #[tokio::test]
+    async fn member_with_editor_role_allowed() {
+        let files = svc_with(
+            Arc::new(EditorAuthzRepo),
+            Arc::new(CapturingEmitter::default()),
+        );
+        let cx = member_cx(AppId::new());
+        files.create(&cx, "c", new_file("f", b"x")).await.unwrap();
+    }
+
+    #[tokio::test]
+    async fn mutations_emit_events_with_correct_prev() {
+        let emitter = Arc::new(CapturingEmitter::default());
+        let files = svc_with(Arc::new(DenyingAuthzRepo), emitter.clone());
+        let cx = anon_cx(AppId::new());
+
+        let id = files.create(&cx, "c", new_file("f", b"one")).await.unwrap();
+        files
+            .update(
+                &cx,
+                "c",
+                &id.to_string(),
+                FileUpdate {
+                    data: b"two".to_vec(),
+                    name: None,
+                    content_type: None,
+                },
+            )
+            .await
+            .unwrap();
+        files.delete(&cx, "c", &id.to_string()).await.unwrap();
+
+        let events = emitter.events.lock().unwrap();
+        assert_eq!(events.len(), 3);
+        // create: prev is None
+        assert_eq!(events[0].op, "create");
+        assert_eq!(events[0].source, "files");
+        assert!(events[0].old_payload.is_none());
+        assert!(events[0].payload.is_some());
+        // update: prev is the prior metadata
+        assert_eq!(events[1].op, "update");
+        assert!(events[1].old_payload.is_some());
+        // delete: prev is the deleted metadata (payload == old_payload)
+        assert_eq!(events[2].op, "delete");
+        assert_eq!(events[2].payload, events[2].old_payload);
+        assert!(events[2].payload.is_some());
+    }
+
+    #[tokio::test]
+    async fn list_cursor_paginates() {
+        let files = svc();
+        let cx = anon_cx(AppId::new());
+        for i in 0..5 {
+            files
+                .create(&cx, "c", new_file(&format!("f{i}"), b"x"))
+                .await
+                .unwrap();
+        }
+        let p1 = files.list(&cx, "c", None, 2).await.unwrap();
+        assert_eq!(p1.files.len(), 2);
+        assert!(p1.next_cursor.is_some());
+        let p2 = files
+            .list(&cx, "c", p1.next_cursor.as_deref(), 2)
+            .await
+            .unwrap();
+        assert_eq!(p2.files.len(), 2);
+        let p3 = files
+            .list(&cx, "c", p2.next_cursor.as_deref(), 2)
+            .await
+            .unwrap();
+        assert_eq!(p3.files.len(), 1);
+        assert!(p3.next_cursor.is_none());
+    }
+}

crates/manager-core/src/http_service.rs

@@ -0,0 +1,793 @@
+//! `HttpServiceImpl` — reqwest-backed outbound HTTP for the v1.1.4
+//! `http::*` SDK.
+//!
+//! Mirrors the v1.1.1+ stateful-service shape (`KvServiceImpl`):
+//! script-as-gate authz (`AppHttpRequest`, skipped when
+//! `cx.principal` is `None`), with the backend talking to the network
+//! instead of Postgres. The reqwest client is built once at startup
+//! with the [`crate::ssrf::SsrfResolver`] wired in via
+//! `dns_resolver`, so the SSRF deny-list applies at every connection —
+//! including each redirect hop, since redirects are followed manually
+//! through the same client.
+//!
+//! Layering vs the raw client:
+//!   1. URL validation: scheme must be http/https; ports 22/25/465/587
+//!      are blocked. (IP-level filtering is the resolver's job.)
+//!   2. Body-size caps on both request and response (stream-with-cap on
+//!      the response, checking `Content-Length` first).
+//!   3. Total-request timeout (default 30s, max 60s) on top of the
+//!      client's 10s connect timeout.
+//!   4. Default `User-Agent` unless the caller set one.
+//!
+//! Bodies/headers are never logged (PII): only url + status + duration
+//! at debug level.
+
+use std::collections::BTreeMap;
+use std::env;
+use std::sync::Arc;
+use std::time::Duration;
+
+use async_trait::async_trait;
+use picloud_shared::{HttpError, HttpRequest, HttpResponse, HttpService, SdkCallCx};
+use reqwest::header::{HeaderMap, HeaderName, HeaderValue, CONTENT_TYPE, LOCATION, USER_AGENT};
+use reqwest::{Client, Method, StatusCode};
+
+use crate::authz::{self, AuthzRepo, Capability};
+use crate::ssrf::{self, SsrfPolicy, SSRF_BLOCK_PREFIX};
+
+/// Default per-request timeout (ms) when the script omits `timeout_ms`.
+pub const DEFAULT_TIMEOUT_MS: u32 = 30_000;
+/// Hard ceiling on the per-request timeout. Values above this are
+/// rejected by the bridge (not silently clamped).
+pub const MAX_TIMEOUT_MS: u32 = 60_000;
+/// Default redirect cap.
+pub const DEFAULT_MAX_REDIRECTS: u32 = 5;
+/// Hard ceiling on redirects.
+pub const MAX_REDIRECTS_CEILING: u32 = 10;
+/// 10 MB default body cap on both directions.
+const DEFAULT_BODY_LIMIT_BYTES: usize = 10 * 1024 * 1024;
+/// DNS + connect + TLS hard cap.
+const CONNECT_TIMEOUT: Duration = Duration::from_secs(10);
+
+/// Outbound-HTTP tunables. Env-overridable following the same pattern
+/// as `TriggerConfig::from_env`.
+#[derive(Debug, Clone, Copy)]
+pub struct HttpConfig {
+    /// Disables the SSRF deny-list entirely. Dev/test only — the binary
+    /// logs a startup warning when this is set.
+    pub allow_private: bool,
+    pub max_request_body_bytes: usize,
+    pub max_response_body_bytes: usize,
+}
+
+impl HttpConfig {
+    #[must_use]
+    pub const fn conservative() -> Self {
+        Self {
+            allow_private: false,
+            max_request_body_bytes: DEFAULT_BODY_LIMIT_BYTES,
+            max_response_body_bytes: DEFAULT_BODY_LIMIT_BYTES,
+        }
+    }
+
+    #[must_use]
+    pub fn from_env() -> Self {
+        let mut c = Self::conservative();
+        if let Ok(v) = env::var("PICLOUD_HTTP_ALLOW_PRIVATE") {
+            c.allow_private =
+                matches!(v.trim().to_ascii_lowercase().as_str(), "1" | "true" | "yes");
+        }
+        load_usize(
+            &mut c.max_request_body_bytes,
+            "PICLOUD_HTTP_MAX_REQUEST_BODY_BYTES",
+        );
+        load_usize(
+            &mut c.max_response_body_bytes,
+            "PICLOUD_HTTP_MAX_RESPONSE_BODY_BYTES",
+        );
+        c
+    }
+}
+
+impl Default for HttpConfig {
+    fn default() -> Self {
+        Self::conservative()
+    }
+}
+
+fn load_usize(dst: &mut usize, key: &str) {
+    if let Ok(v) = env::var(key) {
+        match v.parse::<usize>() {
+            Ok(n) => *dst = n,
+            Err(e) => {
+                tracing::warn!(env = key, error = %e, "ignoring invalid http-config value");
+            }
+        }
+    }
+}
+
+pub struct HttpServiceImpl {
+    client: Client,
+    authz: Arc<dyn AuthzRepo>,
+    config: HttpConfig,
+    /// Same policy wired into the DNS resolver. Held here too because
+    /// reqwest only routes *hostnames* through the custom resolver — a
+    /// URL with a **literal IP** host bypasses it, so literal IPs are
+    /// checked directly at URL-validation time.
+    policy: SsrfPolicy,
+}
+
+impl HttpServiceImpl {
+    /// Build the service, constructing the reqwest client with the SSRF
+    /// resolver. Redirects are followed manually (so per-request limits
+    /// are honored and every hop re-resolves through the SSRF
+    /// resolver), hence `redirect(Policy::none())`.
+    ///
+    /// # Panics
+    ///
+    /// Panics if the reqwest client fails to build — this is a
+    /// startup-time invariant, not a runtime path.
+    #[must_use]
+    pub fn new(config: HttpConfig, authz: Arc<dyn AuthzRepo>) -> Self {
+        let policy = SsrfPolicy::new(config.allow_private);
+        let client = Client::builder()
+            .dns_resolver(ssrf::resolver(policy))
+            .connect_timeout(CONNECT_TIMEOUT)
+            .redirect(reqwest::redirect::Policy::none())
+            .build()
+            .expect("build outbound http client");
+        Self {
+            client,
+            authz,
+            config,
+            policy,
+        }
+    }
+
+    async fn check_request(&self, cx: &SdkCallCx) -> Result<(), HttpError> {
+        if let Some(ref principal) = cx.principal {
+            authz::require(
+                &*self.authz,
+                principal,
+                Capability::AppHttpRequest(cx.app_id),
+            )
+            .await
+            .map_err(|_| HttpError::Forbidden)?;
+        }
+        Ok(())
+    }
+}
+
+#[async_trait]
+impl HttpService for HttpServiceImpl {
+    async fn request(&self, cx: &SdkCallCx, req: HttpRequest) -> Result<HttpResponse, HttpError> {
+        self.check_request(cx).await?;
+
+        // Request body cap.
+        if let Some(ref body) = req.body {
+            if body.len() > self.config.max_request_body_bytes {
+                return Err(HttpError::BodyTooLarge("request"));
+            }
+        }
+
+        let timeout = Duration::from_millis(u64::from(req.timeout_ms.min(MAX_TIMEOUT_MS)));
+        let started = std::time::Instant::now();
+        let url_for_log = req.url.clone();
+
+        // Whole-request budget (DNS + connect + TLS + all redirect hops
+        // + body read). Connect alone is further bounded by the
+        // client's CONNECT_TIMEOUT.
+        let outcome = match tokio::time::timeout(timeout, self.run(req)).await {
+            Ok(r) => r,
+            Err(_) => Err(HttpError::Timeout),
+        };
+
+        let duration_ms = u64::try_from(started.elapsed().as_millis()).unwrap_or(u64::MAX);
+        match &outcome {
+            Ok(resp) => tracing::debug!(
+                url = %url_for_log,
+                status = resp.status,
+                duration_ms,
+                "outbound http"
+            ),
+            Err(err) => tracing::debug!(
+                url = %url_for_log,
+                error = %err,
+                duration_ms,
+                "outbound http failed"
+            ),
+        }
+        outcome
+    }
+}
+
+impl HttpServiceImpl {
+    /// Core request path: validate, build headers, follow redirects
+    /// manually, read the response body with a cap.
+    async fn run(&self, req: HttpRequest) -> Result<HttpResponse, HttpError> {
+        let method = Method::from_bytes(req.method.as_bytes())
+            .map_err(|_| HttpError::Backend(format!("invalid method: {}", req.method)))?;
+
+        let mut current = url::Url::parse(&req.url)
+            .map_err(|e| HttpError::InvalidUrl(format!("{}: {e}", req.url)))?;
+        validate_url(&current, self.policy)?;
+
+        let mut header_map = build_headers(&req, &current)?;
+        let mut method = method;
+        let mut body = req.body.clone();
+        let mut redirects: u32 = 0;
+        let max_redirects = req.max_redirects.min(MAX_REDIRECTS_CEILING);
+
+        loop {
+            // Re-validate scheme/port (and literal-IP SSRF) on each hop.
+            // Hostname IP filtering is the resolver's job and runs
+            // automatically at connect time.
+            validate_url(&current, self.policy)?;
+
+            let mut rb = self.client.request(method.clone(), current.clone());
+            rb = rb.headers(header_map.clone());
+            if let Some(ref b) = body {
+                rb = rb.body(b.clone());
+            }
+            let resp = rb.send().await.map_err(map_reqwest_err)?;
+            let status = resp.status();
+
+            if req.follow_redirects && is_redirect(status) {
+                if let Some(loc) = resp.headers().get(LOCATION) {
+                    if redirects >= max_redirects {
+                        return Err(HttpError::Backend(format!(
+                            "too many redirects (max {max_redirects})"
+                        )));
+                    }
+                    redirects += 1;
+                    let loc_str = loc.to_str().map_err(|_| {
+                        HttpError::Backend("redirect Location not valid UTF-8".into())
+                    })?;
+                    current = current
+                        .join(loc_str)
+                        .map_err(|e| HttpError::InvalidUrl(format!("redirect target: {e}")))?;
+
+                    // 303 always → GET; 301/302 historically downgrade
+                    // POST→GET (matches browsers). 307/308 preserve.
+                    if matches!(status.as_u16(), 301..=303) {
+                        method = Method::GET;
+                        body = None;
+                        header_map.remove(CONTENT_TYPE);
+                    }
+                    continue;
+                }
+            }
+
+            return self.read_capped(resp).await;
+        }
+    }
+
+    async fn read_capped(&self, resp: reqwest::Response) -> Result<HttpResponse, HttpError> {
+        let status = resp.status().as_u16();
+        let mut headers = BTreeMap::new();
+        for (name, value) in resp.headers() {
+            // Header names lowercased per the documented response shape.
+            headers.insert(
+                name.as_str().to_ascii_lowercase(),
+                value.to_str().unwrap_or("").to_string(),
+            );
+        }
+
+        let cap = self.config.max_response_body_bytes;
+        if let Some(len) = resp.content_length() {
+            if len > cap as u64 {
+                return Err(HttpError::BodyTooLarge("response"));
+            }
+        }
+
+        let mut buf: Vec<u8> = Vec::new();
+        let mut resp = resp;
+        while let Some(chunk) = resp.chunk().await.map_err(map_reqwest_err)? {
+            if buf.len() + chunk.len() > cap {
+                return Err(HttpError::BodyTooLarge("response"));
+            }
+            buf.extend_from_slice(&chunk);
+        }
+        let body_raw = String::from_utf8_lossy(&buf).into_owned();
+        Ok(HttpResponse {
+            status,
+            headers,
+            body_raw,
+        })
+    }
+}
+
+/// http/https only; block the SSH + SMTP ports; apply the SSRF policy
+/// to **literal-IP** hosts (hostnames are filtered by the DNS resolver
+/// at connect time, but literal IPs never reach the resolver).
+fn validate_url(url: &url::Url, policy: SsrfPolicy) -> Result<(), HttpError> {
+    match url.scheme() {
+        "http" | "https" => {}
+        other => return Err(HttpError::BlockedScheme(other.to_string())),
+    }
+    match url.host() {
+        None => return Err(HttpError::InvalidUrl("missing host".into())),
+        Some(url::Host::Ipv4(ip)) => {
+            policy
+                .check(std::net::IpAddr::V4(ip))
+                .map_err(|reason| HttpError::Ssrf(reason.to_string()))?;
+        }
+        Some(url::Host::Ipv6(ip)) => {
+            policy
+                .check(std::net::IpAddr::V6(ip))
+                .map_err(|reason| HttpError::Ssrf(reason.to_string()))?;
+        }
+        Some(url::Host::Domain(_)) => {}
+    }
+    let port = url
+        .port_or_known_default()
+        .unwrap_or(if url.scheme() == "https" { 443 } else { 80 });
+    if matches!(port, 22 | 25 | 465 | 587) {
+        return Err(HttpError::BlockedPort(port));
+    }
+    Ok(())
+}
+
+/// Build the request header map: merge caller headers, then apply the
+/// default `User-Agent` (unless overridden) and the bridge-chosen
+/// `Content-Type` (unless overridden).
+fn build_headers(req: &HttpRequest, _url: &url::Url) -> Result<HeaderMap, HttpError> {
+    let mut map = HeaderMap::new();
+    let mut has_user_agent = false;
+    let mut has_content_type = false;
+    for (k, v) in &req.headers {
+        let name = HeaderName::from_bytes(k.as_bytes())
+            .map_err(|_| HttpError::Backend(format!("invalid header name: {k}")))?;
+        let value = HeaderValue::from_str(v)
+            .map_err(|_| HttpError::Backend(format!("invalid header value for {k}")))?;
+        if name == USER_AGENT {
+            has_user_agent = true;
+        }
+        if name == CONTENT_TYPE {
+            has_content_type = true;
+        }
+        map.append(name, value);
+    }
+
+    if !has_user_agent {
+        let script = req.script_id.as_deref().unwrap_or("unknown");
+        let ua = format!(
+            "picloud/{} (script:{})",
+            picloud_shared::PRODUCT_VERSION,
+            script
+        );
+        if let Ok(value) = HeaderValue::from_str(&ua) {
+            map.insert(USER_AGENT, value);
+        }
+    }
+
+    if !has_content_type {
+        if let Some(ref ct) = req.content_type {
+            if let Ok(value) = HeaderValue::from_str(ct) {
+                map.insert(CONTENT_TYPE, value);
+            }
+        }
+    }
+
+    Ok(map)
+}
+
+const fn is_redirect(status: StatusCode) -> bool {
+    matches!(status.as_u16(), 301..=303 | 307 | 308)
+}
+
+/// Map a reqwest error to an `HttpError`, never leaking the resolved
+/// IP. SSRF blocks are detected by scanning the error source chain for
+/// the resolver's marker prefix.
+fn map_reqwest_err(err: reqwest::Error) -> HttpError {
+    if let Some(reason) = ssrf_reason(&err) {
+        return HttpError::Ssrf(reason);
+    }
+    if err.is_timeout() {
+        return HttpError::Timeout;
+    }
+    if err.is_connect() {
+        return HttpError::Network("connection failed".into());
+    }
+    if err.is_request() {
+        return HttpError::Network("request failed".into());
+    }
+    HttpError::Network("network error".into())
+}
+
+/// Walk the error source chain looking for the SSRF marker the resolver
+/// embeds. Returns the category reason (no IP) when found.
+fn ssrf_reason(err: &reqwest::Error) -> Option<String> {
+    let mut src: Option<&(dyn std::error::Error + 'static)> = Some(err);
+    while let Some(e) = src {
+        let s = e.to_string();
+        if let Some(idx) = s.find(SSRF_BLOCK_PREFIX) {
+            return Some(s[idx + SSRF_BLOCK_PREFIX.len()..].to_string());
+        }
+        src = e.source();
+    }
+    None
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::authz::AuthzError;
+    use async_trait::async_trait;
+    use picloud_shared::{
+        AdminUserId, AppId, AppRole, ExecutionId, InstanceRole, Principal, RequestId, ScriptId,
+        UserId,
+    };
+    use std::collections::BTreeMap;
+    use std::io::Write as _;
+    use std::net::SocketAddr;
+    use tokio::io::{AsyncReadExt, AsyncWriteExt};
+    use tokio::net::TcpListener;
+
+    struct AllowAuthz;
+    #[async_trait]
+    impl AuthzRepo for AllowAuthz {
+        async fn membership(&self, _u: UserId, _a: AppId) -> Result<Option<AppRole>, AuthzError> {
+            Ok(Some(AppRole::Editor))
+        }
+    }
+    struct DenyAuthz;
+    #[async_trait]
+    impl AuthzRepo for DenyAuthz {
+        async fn membership(&self, _u: UserId, _a: AppId) -> Result<Option<AppRole>, AuthzError> {
+            Ok(None)
+        }
+    }
+
+    fn dev_service(authz: Arc<dyn AuthzRepo>) -> HttpServiceImpl {
+        // allow_private so the test TcpListener on 127.0.0.1 is reachable.
+        let mut config = HttpConfig::conservative();
+        config.allow_private = true;
+        HttpServiceImpl::new(config, authz)
+    }
+
+    fn anon_cx() -> SdkCallCx {
+        SdkCallCx {
+            app_id: AppId::new(),
+            script_id: ScriptId::new(),
+            principal: None,
+            execution_id: ExecutionId::new(),
+            request_id: RequestId::new(),
+            trigger_depth: 0,
+            root_execution_id: ExecutionId::new(),
+            is_dead_letter_handler: false,
+            event: None,
+        }
+    }
+
+    fn member_cx() -> SdkCallCx {
+        let mut cx = anon_cx();
+        cx.principal = Some(Principal {
+            user_id: AdminUserId::new(),
+            instance_role: InstanceRole::Member,
+            scopes: None,
+            app_binding: None,
+        });
+        cx
+    }
+
+    fn req(method: &str, url: String) -> HttpRequest {
+        HttpRequest {
+            method: method.into(),
+            url,
+            headers: BTreeMap::new(),
+            body: None,
+            content_type: None,
+            timeout_ms: 5000,
+            follow_redirects: true,
+            max_redirects: 5,
+            script_id: Some("test-script".into()),
+        }
+    }
+
+    /// Minimal single-shot HTTP/1.1 server. Reads the request, runs
+    /// `handler` to produce the raw response bytes, writes them, closes.
+    /// Returns the bound address.
+    async fn spawn_server<F>(handler: F) -> SocketAddr
+    where
+        F: Fn(String) -> Vec<u8> + Send + Sync + 'static,
+    {
+        let listener = TcpListener::bind("127.0.0.1:0").await.unwrap();
+        let addr = listener.local_addr().unwrap();
+        tokio::spawn(async move {
+            loop {
+                let Ok((mut sock, _)) = listener.accept().await else {
+                    break;
+                };
+                let mut buf = vec![0u8; 65536];
+                let n = sock.read(&mut buf).await.unwrap_or(0);
+                let request = String::from_utf8_lossy(&buf[..n]).to_string();
+                let response = handler(request);
+                let _ = sock.write_all(&response).await;
+                let _ = sock.flush().await;
+            }
+        });
+        addr
+    }
+
+    fn ok_response(body: &str, content_type: &str) -> Vec<u8> {
+        let mut v = Vec::new();
+        write!(
+            v,
+            "HTTP/1.1 200 OK\r\nContent-Type: {content_type}\r\nContent-Length: {}\r\nConnection: close\r\n\r\n{body}",
+            body.len()
+        )
+        .unwrap();
+        v
+    }
+
+    #[tokio::test]
+    async fn get_round_trip() {
+        let addr = spawn_server(|_req| ok_response("hello", "text/plain")).await;
+        let svc = dev_service(Arc::new(AllowAuthz));
+        let resp = svc
+            .request(&anon_cx(), req("GET", format!("http://{addr}/")))
+            .await
+            .unwrap();
+        assert_eq!(resp.status, 200);
+        assert_eq!(resp.body_raw, "hello");
+        assert_eq!(
+            resp.headers.get("content-type").map(String::as_str),
+            Some("text/plain")
+        );
+    }
+
+    #[tokio::test]
+    async fn post_sends_body_and_default_user_agent() {
+        let addr = spawn_server(|request| {
+            // Echo back whether the body + default UA were present.
+            let has_ua = request.to_lowercase().contains("user-agent: picloud/");
+            let has_body = request.contains("xyzzy");
+            ok_response(&format!("ua={has_ua},body={has_body}"), "text/plain")
+        })
+        .await;
+        let svc = dev_service(Arc::new(AllowAuthz));
+        let mut r = req("POST", format!("http://{addr}/"));
+        r.body = Some(b"xyzzy".to_vec());
+        r.content_type = Some("text/plain".into());
+        let resp = svc.request(&anon_cx(), r).await.unwrap();
+        assert_eq!(resp.body_raw, "ua=true,body=true");
+    }
+
+    #[tokio::test]
+    async fn custom_user_agent_overrides_default() {
+        let addr = spawn_server(|request| {
+            let has_custom = request.to_lowercase().contains("user-agent: my-agent");
+            let has_default = request.to_lowercase().contains("picloud/");
+            ok_response(
+                &format!("custom={has_custom},default={has_default}"),
+                "text/plain",
+            )
+        })
+        .await;
+        let svc = dev_service(Arc::new(AllowAuthz));
+        let mut r = req("GET", format!("http://{addr}/"));
+        r.headers.insert("User-Agent".into(), "my-agent".into());
+        let resp = svc.request(&anon_cx(), r).await.unwrap();
+        assert_eq!(resp.body_raw, "custom=true,default=false");
+    }
+
+    #[tokio::test]
+    async fn empty_body_response() {
+        let addr = spawn_server(|_r| {
+            b"HTTP/1.1 204 No Content\r\nContent-Length: 0\r\nConnection: close\r\n\r\n".to_vec()
+        })
+        .await;
+        let svc = dev_service(Arc::new(AllowAuthz));
+        let resp = svc
+            .request(&anon_cx(), req("GET", format!("http://{addr}/")))
+            .await
+            .unwrap();
+        assert_eq!(resp.status, 204);
+        assert_eq!(resp.body_raw, "");
+    }
+
+    #[tokio::test]
+    async fn non_2xx_does_not_error() {
+        let addr = spawn_server(|_r| {
+            b"HTTP/1.1 500 Internal Server Error\r\nContent-Length: 3\r\nConnection: close\r\n\r\nerr".to_vec()
+        })
+        .await;
+        let svc = dev_service(Arc::new(AllowAuthz));
+        let resp = svc
+            .request(&anon_cx(), req("GET", format!("http://{addr}/")))
+            .await
+            .unwrap();
+        assert_eq!(resp.status, 500);
+        assert_eq!(resp.body_raw, "err");
+    }
+
+    #[tokio::test]
+    async fn response_over_content_length_cap_rejected() {
+        let addr = spawn_server(|_r| ok_response("0123456789", "text/plain")).await;
+        let mut config = HttpConfig::conservative();
+        config.allow_private = true;
+        config.max_response_body_bytes = 5; // body is 10 bytes
+        let svc = HttpServiceImpl::new(config, Arc::new(AllowAuthz));
+        let err = svc
+            .request(&anon_cx(), req("GET", format!("http://{addr}/")))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, HttpError::BodyTooLarge("response")));
+    }
+
+    #[tokio::test]
+    async fn response_over_cap_without_content_length_rejected_mid_stream() {
+        // No Content-Length header → must be caught while streaming.
+        let addr = spawn_server(|_r| {
+            b"HTTP/1.1 200 OK\r\nConnection: close\r\n\r\n0123456789ABCDEF".to_vec()
+        })
+        .await;
+        let mut config = HttpConfig::conservative();
+        config.allow_private = true;
+        config.max_response_body_bytes = 4;
+        let svc = HttpServiceImpl::new(config, Arc::new(AllowAuthz));
+        let err = svc
+            .request(&anon_cx(), req("GET", format!("http://{addr}/")))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, HttpError::BodyTooLarge("response")));
+    }
+
+    #[tokio::test]
+    async fn request_body_over_cap_rejected_before_send() {
+        let mut config = HttpConfig::conservative();
+        config.allow_private = true;
+        config.max_request_body_bytes = 3;
+        let svc = HttpServiceImpl::new(config, Arc::new(AllowAuthz));
+        let mut r = req("POST", "http://127.0.0.1:1/".into());
+        r.body = Some(b"too long".to_vec());
+        let err = svc.request(&anon_cx(), r).await.unwrap_err();
+        assert!(matches!(err, HttpError::BodyTooLarge("request")));
+    }
+
+    #[tokio::test]
+    async fn redirect_followed_up_to_then_throws_beyond_max() {
+        // Server always 302s to itself → unbounded redirect loop,
+        // bounded by max_redirects.
+        let listener = TcpListener::bind("127.0.0.1:0").await.unwrap();
+        let addr = listener.local_addr().unwrap();
+        tokio::spawn(async move {
+            loop {
+                let Ok((mut sock, _)) = listener.accept().await else {
+                    break;
+                };
+                let mut buf = vec![0u8; 4096];
+                let _ = sock.read(&mut buf).await;
+                let body = format!(
+                    "HTTP/1.1 302 Found\r\nLocation: http://{addr}/next\r\nContent-Length: 0\r\nConnection: close\r\n\r\n"
+                );
+                let _ = sock.write_all(body.as_bytes()).await;
+            }
+        });
+        let svc = dev_service(Arc::new(AllowAuthz));
+        let mut r = req("GET", format!("http://{addr}/"));
+        r.max_redirects = 2;
+        let err = svc.request(&anon_cx(), r).await.unwrap_err();
+        assert!(
+            matches!(err, HttpError::Backend(ref m) if m.contains("too many redirects")),
+            "expected too-many-redirects, got {err:?}"
+        );
+    }
+
+    #[tokio::test]
+    async fn scheme_rejected() {
+        let svc = dev_service(Arc::new(AllowAuthz));
+        for url in ["file:///etc/passwd", "ftp://host/x", "gopher://host/"] {
+            let err = svc
+                .request(&anon_cx(), req("GET", url.into()))
+                .await
+                .unwrap_err();
+            match err {
+                HttpError::BlockedScheme(s) => {
+                    assert!(url.starts_with(&s), "scheme {s} not in url {url}");
+                }
+                other => panic!("expected BlockedScheme for {url}, got {other:?}"),
+            }
+        }
+    }
+
+    #[tokio::test]
+    async fn ports_rejected() {
+        let svc = dev_service(Arc::new(AllowAuthz));
+        for port in [22u16, 25, 465, 587] {
+            let err = svc
+                .request(
+                    &anon_cx(),
+                    req("GET", format!("http://example.com:{port}/")),
+                )
+                .await
+                .unwrap_err();
+            assert!(
+                matches!(err, HttpError::BlockedPort(p) if p == port),
+                "port {port} should be blocked, got {err:?}"
+            );
+        }
+    }
+
+    #[tokio::test]
+    async fn ssrf_blocks_loopback_without_allow_private() {
+        // Default config (deny-list ON). A request to a loopback host
+        // must surface as Ssrf, not a generic network error.
+        let svc = HttpServiceImpl::new(HttpConfig::conservative(), Arc::new(AllowAuthz));
+        let err = svc
+            .request(&anon_cx(), req("GET", "http://127.0.0.1:9/".into()))
+            .await
+            .unwrap_err();
+        match err {
+            HttpError::Ssrf(reason) => {
+                assert_eq!(reason, "loopback");
+                assert!(!reason.contains("127.0.0.1"), "reason must not leak the IP");
+            }
+            other => panic!("expected Ssrf, got {other:?}"),
+        }
+    }
+
+    #[tokio::test]
+    async fn ssrf_blocks_hostname_resolving_to_loopback() {
+        // `localhost` resolves to 127.0.0.1 / ::1 — all denied. This
+        // exercises the DNS-resolver path (vs the literal-IP path) and
+        // must surface as Ssrf, not a generic DNS error.
+        let svc = HttpServiceImpl::new(HttpConfig::conservative(), Arc::new(AllowAuthz));
+        let err = svc
+            .request(&anon_cx(), req("GET", "http://localhost:9/".into()))
+            .await
+            .unwrap_err();
+        assert!(
+            matches!(err, HttpError::Ssrf(_)),
+            "expected Ssrf for localhost, got {err:?}"
+        );
+    }
+
+    #[tokio::test]
+    async fn timeout_throws() {
+        // Server that accepts then never responds.
+        let listener = TcpListener::bind("127.0.0.1:0").await.unwrap();
+        let addr = listener.local_addr().unwrap();
+        tokio::spawn(async move {
+            if let Ok((sock, _)) = listener.accept().await {
+                // Hold the socket open without replying.
+                tokio::time::sleep(Duration::from_secs(30)).await;
+                drop(sock);
+            }
+        });
+        let svc = dev_service(Arc::new(AllowAuthz));
+        let mut r = req("GET", format!("http://{addr}/"));
+        r.timeout_ms = 300;
+        let err = svc.request(&anon_cx(), r).await.unwrap_err();
+        assert!(matches!(err, HttpError::Timeout), "got {err:?}");
+    }
+
+    #[tokio::test]
+    async fn anon_skips_authz_member_without_scope_forbidden() {
+        let addr = spawn_server(|_r| ok_response("ok", "text/plain")).await;
+        // Anonymous principal → authz skipped even with DenyAuthz.
+        let svc = dev_service(Arc::new(DenyAuthz));
+        let ok = svc
+            .request(&anon_cx(), req("GET", format!("http://{addr}/")))
+            .await;
+        assert!(ok.is_ok());
+        // Authenticated member with no role → Forbidden.
+        let err = svc
+            .request(&member_cx(), req("GET", format!("http://{addr}/")))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, HttpError::Forbidden));
+    }
+
+    #[tokio::test]
+    async fn member_with_role_allowed() {
+        let addr = spawn_server(|_r| ok_response("ok", "text/plain")).await;
+        let svc = dev_service(Arc::new(AllowAuthz));
+        let resp = svc
+            .request(&member_cx(), req("GET", format!("http://{addr}/")))
+            .await
+            .unwrap();
+        assert_eq!(resp.status, 200);
+    }
+}

crates/manager-core/src/kv_service.rs

@@ -188,7 +188,7 @@ mod tests {
     use async_trait::async_trait;
     use picloud_shared::{
         AdminUserId, AppId, AppRole, ExecutionId, InstanceRole, NoopEventEmitter, Principal,
-        RequestId, UserId,
+        RequestId, ScriptId, UserId,
     };
     use std::collections::{BTreeMap, HashMap};
     use tokio::sync::Mutex;
@@ -301,6 +301,7 @@ mod tests {
     fn anon_cx(app_id: AppId) -> SdkCallCx {
         SdkCallCx {
             app_id,
+            script_id: ScriptId::new(),
             principal: None,
             execution_id: ExecutionId::new(),
             request_id: RequestId::new(),
@@ -314,6 +315,7 @@ mod tests {
     fn owner_cx(app_id: AppId) -> SdkCallCx {
         SdkCallCx {
             app_id,
+            script_id: ScriptId::new(),
             principal: Some(Principal {
                 user_id: AdminUserId::new(),
                 instance_role: InstanceRole::Owner,
@@ -332,6 +334,7 @@ mod tests {
     fn member_no_role_cx(app_id: AppId) -> SdkCallCx {
         SdkCallCx {
             app_id,
+            script_id: ScriptId::new(),
             principal: Some(Principal {
                 user_id: AdminUserId::new(),
                 instance_role: InstanceRole::Member,

crates/manager-core/src/lib.rs

@@ -22,23 +22,35 @@ pub mod auth_api;
 pub mod auth_bootstrap;
 pub mod auth_middleware;
 pub mod authz;
+pub mod cron_scheduler;
 pub mod dead_letter_repo;
 pub mod dead_letter_service;
 pub mod dead_letters_api;
 pub mod dispatcher;
+pub mod docs_filter;
+pub mod docs_repo;
+pub mod docs_service;
+pub mod files_api;
+pub mod files_repo;
+pub mod files_service;
 pub mod gc;
+pub mod http_service;
 pub mod kv_repo;
 pub mod kv_service;
 pub mod log_sink;
 pub mod migrations;
+pub mod module_source;
 pub mod outbox_event_emitter;
 pub mod outbox_repo;
 pub mod principal_resolver;
+pub mod pubsub_repo;
+pub mod pubsub_service;
 pub mod repo;
 pub mod route_admin;
 pub mod route_repo;
 pub mod sandbox;
 pub mod scheduler;
+pub mod ssrf;
 pub mod trigger_config;
 pub mod trigger_repo;
 pub mod triggers_api;
@@ -80,21 +92,31 @@ pub use auth_middleware::{
     API_KEY_PREFIX, API_KEY_PREFIX_LEN, SESSION_COOKIE,
 };
 pub use authz::{can, require, AuthzDenied, AuthzError, AuthzRepo, Capability, Decision};
+pub use cron_scheduler::spawn_cron_scheduler;
 pub use dead_letter_repo::{
     DeadLetterRepo, DeadLetterRepoError, DeadLetterRow, NewDeadLetter, PostgresDeadLetterRepo,
 };
 pub use dead_letter_service::PostgresDeadLetterService;
 pub use dead_letters_api::{dead_letters_router, DeadLettersApiError, DeadLettersState};
 pub use dispatcher::{compute_backoff, Dispatcher, DispatcherError};
+pub use docs_repo::{DocsRepo, DocsRepoError, PostgresDocsRepo};
+pub use docs_service::DocsServiceImpl;
+pub use files_api::{files_admin_router, FilesAdminState};
+pub use files_repo::{FilesConfig, FilesRepo, FilesRepoError, FsFilesRepo};
+pub use files_service::FilesServiceImpl;
 pub use gc::{spawn_abandoned_gc, spawn_dead_letter_gc};
+pub use http_service::{HttpConfig, HttpServiceImpl};
 pub use kv_repo::{KvRepo, KvRepoError, PostgresKvRepo};
 pub use kv_service::KvServiceImpl;
 pub use log_sink::PostgresExecutionLogSink;
+pub use module_source::PostgresModuleSource;
 pub use outbox_event_emitter::OutboxEventEmitter;
 pub use outbox_repo::{
     NewOutboxRow, OutboxRepo, OutboxRepoError, OutboxRow, OutboxSourceKind, PostgresOutboxRepo,
 };
 pub use principal_resolver::{AdminPrincipalResolver, PrincipalResolver, PrincipalResolverError};
+pub use pubsub_repo::{PostgresPubsubRepo, PublishCtx, PubsubRepo, PubsubRepoError};
+pub use pubsub_service::PubsubServiceImpl;
 pub use repo::{
     ExecutionLogRepository, NewScript, PostgresExecutionLogRepository, PostgresScriptRepository,
     RepoResolver, ScriptPatch, ScriptRepository, ScriptRepositoryError,
@@ -104,8 +126,9 @@ pub use route_repo::{NewRoute, PostgresRouteRepository, RouteRepository};
 pub use sandbox::{CeilingError, SandboxCeiling};
 pub use trigger_config::{BackoffShape, TriggerConfig};
 pub use trigger_repo::{
-    collection_matches, CreateDeadLetterTrigger, CreateKvTrigger, DeadLetterTriggerMatch,
-    KvTriggerMatch, PostgresTriggerRepo, Trigger, TriggerDetails, TriggerDispatchMode, TriggerKind,
-    TriggerRepo, TriggerRepoError,
+    collection_matches, CreateDeadLetterTrigger, CreateDocsTrigger, CreateFilesTrigger,
+    CreateKvTrigger, CreatePubsubTrigger, DeadLetterTriggerMatch, DocsTriggerMatch,
+    FilesTriggerMatch, KvTriggerMatch, PostgresTriggerRepo, Trigger, TriggerDetails,
+    TriggerDispatchMode, TriggerKind, TriggerRepo, TriggerRepoError,
 };
 pub use triggers_api::{triggers_router, TriggersApiError, TriggersState};

crates/manager-core/src/module_source.rs

@@ -0,0 +1,74 @@
+//! `PostgresModuleSource` — the Postgres-backed `ModuleSource` impl.
+//!
+//! Mirrors the structure of [`crate::kv_repo::PostgresKvRepo`] /
+//! [`crate::docs_repo::PostgresDocsRepo`]: thin wrapper around a
+//! `PgPool` that owns a single statement returning the module by
+//! `(cx.app_id, name, kind = 'module')`. The resolver lives in
+//! `executor-core` and consumes this trait through the `Services`
+//! bundle, so manager-core stays the only crate that touches
+//! Postgres.
+
+use async_trait::async_trait;
+use chrono::{DateTime, Utc};
+use picloud_shared::{ModuleScript, ModuleSource, ModuleSourceError, SdkCallCx};
+use sqlx::PgPool;
+
+pub struct PostgresModuleSource {
+    pool: PgPool,
+}
+
+impl PostgresModuleSource {
+    #[must_use]
+    pub fn new(pool: PgPool) -> Self {
+        Self { pool }
+    }
+}
+
+#[derive(sqlx::FromRow)]
+struct ModuleRow {
+    id: uuid::Uuid,
+    app_id: uuid::Uuid,
+    name: String,
+    source: String,
+    updated_at: DateTime<Utc>,
+}
+
+impl From<ModuleRow> for ModuleScript {
+    fn from(r: ModuleRow) -> Self {
+        Self {
+            script_id: r.id.into(),
+            app_id: r.app_id.into(),
+            name: r.name,
+            source: r.source,
+            updated_at: r.updated_at,
+        }
+    }
+}
+
+#[async_trait]
+impl ModuleSource for PostgresModuleSource {
+    async fn lookup(
+        &self,
+        cx: &SdkCallCx,
+        name: &str,
+    ) -> Result<Option<ModuleScript>, ModuleSourceError> {
+        // The query is the cross-app isolation boundary: app_id comes
+        // from cx (never from the script-passed argument), and the
+        // CHECK constraint `kind IN ('endpoint','module')` plus the
+        // `kind = 'module'` filter together guarantee endpoint scripts
+        // are never importable. The `(app_id, kind)` index from
+        // migration 0015 makes this an index scan returning at most
+        // one row (per-app uniqueness on `name`).
+        let row: Option<ModuleRow> = sqlx::query_as(
+            "SELECT id, app_id, name, source, updated_at \
+             FROM scripts \
+             WHERE app_id = $1 AND kind = 'module' AND name = $2",
+        )
+        .bind(cx.app_id.into_inner())
+        .bind(name)
+        .fetch_optional(&self.pool)
+        .await
+        .map_err(|e| ModuleSourceError::Backend(e.to_string()))?;
+        Ok(row.map(Into::into))
+    }
+}

crates/manager-core/src/outbox_event_emitter.rs

@@ -19,7 +19,8 @@ use std::sync::Arc;
 
 use async_trait::async_trait;
 use picloud_shared::{
-    EmitError, KvEventOp, SdkCallCx, ServiceEvent, ServiceEventEmitter, TriggerEvent,
+    DocsEventOp, EmitError, FileMeta, FilesEventOp, KvEventOp, SdkCallCx, ServiceEvent,
+    ServiceEventEmitter, TriggerEvent,
 };
 
 use crate::outbox_repo::{NewOutboxRow, OutboxRepo, OutboxSourceKind};
@@ -42,6 +43,8 @@ impl ServiceEventEmitter for OutboxEventEmitter {
     async fn emit(&self, cx: &SdkCallCx, event: ServiceEvent) -> Result<(), EmitError> {
         match event.source {
             "kv" => self.emit_kv(cx, event).await,
+            "docs" => self.emit_docs(cx, event).await,
+            "files" => self.emit_files(cx, event).await,
             // Future sources land here. For now, silently drop — the
             // SDK calls `events.emit(...)` unconditionally for forward
             // compat, so swallowing without an error is correct.
@@ -100,4 +103,121 @@ impl OutboxEventEmitter {
         }
         Ok(())
     }
+
+    /// v1.1.2. Mirrors `emit_kv` — fan out a docs mutation across
+    /// matching docs triggers + write one outbox row each. The
+    /// `prev_data` change-data-capture surface is preserved from the
+    /// `ServiceEvent.old_payload` field (set by `DocsServiceImpl` on
+    /// update and delete; `None` for create).
+    async fn emit_docs(&self, cx: &SdkCallCx, event: ServiceEvent) -> Result<(), EmitError> {
+        let Some(op) = DocsEventOp::from_wire(event.op) else {
+            return Ok(());
+        };
+        let Some(collection) = event.collection.clone() else {
+            return Ok(());
+        };
+        let id = event.key.clone().unwrap_or_default();
+
+        let matches = self
+            .triggers
+            .list_matching_docs(cx.app_id, &collection, op)
+            .await
+            .map_err(|e| EmitError::Unavailable(format!("trigger lookup: {e}")))?;
+
+        if matches.is_empty() {
+            return Ok(());
+        }
+
+        let trigger_event = TriggerEvent::Docs {
+            op,
+            collection,
+            id,
+            data: event.payload.clone(),
+            prev_data: event.old_payload.clone(),
+        };
+        let payload = serde_json::to_value(&trigger_event)
+            .map_err(|e| EmitError::Rejected(format!("event serialize: {e}")))?;
+
+        for m in matches {
+            self.outbox
+                .insert(NewOutboxRow {
+                    app_id: cx.app_id,
+                    source_kind: OutboxSourceKind::Docs,
+                    trigger_id: Some(m.trigger_id),
+                    script_id: Some(m.script_id),
+                    reply_to: None,
+                    payload: payload.clone(),
+                    origin_principal: cx.principal.as_ref().map(|p| p.user_id),
+                    trigger_depth: cx.trigger_depth.saturating_add(1),
+                    root_execution_id: Some(cx.root_execution_id),
+                })
+                .await
+                .map_err(|e| EmitError::Unavailable(format!("outbox insert: {e}")))?;
+        }
+        Ok(())
+    }
+
+    /// v1.1.5. Fan out a files mutation across matching files triggers.
+    /// The `ServiceEvent.payload` is the file **metadata** (never the
+    /// blob bytes); `old_payload` is the prior metadata (the deleted
+    /// row's metadata on delete). The `TriggerEvent::Files` carries the
+    /// metadata fields explicitly + `prev` for the change-data-capture
+    /// surface.
+    async fn emit_files(&self, cx: &SdkCallCx, event: ServiceEvent) -> Result<(), EmitError> {
+        let Some(op) = FilesEventOp::from_wire(event.op) else {
+            return Ok(());
+        };
+        let Some(collection) = event.collection.clone() else {
+            return Ok(());
+        };
+        // The payload is the FileMeta JSON the FilesServiceImpl emitted.
+        let Some(meta) = event
+            .payload
+            .clone()
+            .and_then(|v| serde_json::from_value::<FileMeta>(v).ok())
+        else {
+            return Ok(());
+        };
+
+        let matches = self
+            .triggers
+            .list_matching_files(cx.app_id, &collection, op)
+            .await
+            .map_err(|e| EmitError::Unavailable(format!("trigger lookup: {e}")))?;
+
+        if matches.is_empty() {
+            return Ok(());
+        }
+
+        let trigger_event = TriggerEvent::Files {
+            op,
+            collection,
+            id: meta.id.to_string(),
+            name: meta.name,
+            content_type: meta.content_type,
+            size: meta.size,
+            checksum: meta.checksum,
+            prev: event.old_payload.clone(),
+        };
+        let payload = serde_json::to_value(&trigger_event)
+            .map_err(|e| EmitError::Rejected(format!("event serialize: {e}")))?;
+
+        for m in matches {
+            self.outbox
+                .insert(NewOutboxRow {
+                    app_id: cx.app_id,
+                    source_kind: OutboxSourceKind::Files,
+                    trigger_id: Some(m.trigger_id),
+                    script_id: Some(m.script_id),
+                    reply_to: None,
+                    payload: payload.clone(),
+                    origin_principal: cx.principal.as_ref().map(|p| p.user_id),
+                    trigger_depth: cx.trigger_depth.saturating_add(1),
+                    root_execution_id: Some(cx.root_execution_id),
+                })
+                .await
+                .map_err(|e| EmitError::Unavailable(format!("outbox insert: {e}")))?;
+        }
+        Ok(())
+    }
 }

crates/manager-core/src/outbox_repo.rs

@@ -22,7 +22,15 @@ pub enum OutboxRepoError {
 pub enum OutboxSourceKind {
     Http,
     Kv,
+    /// v1.1.2.
+    Docs,
     DeadLetter,
+    /// v1.1.4.
+    Cron,
+    /// v1.1.5.
+    Files,
+    /// v1.1.5.
+    Pubsub,
 }
 
 impl OutboxSourceKind {
@@ -31,7 +39,11 @@ impl OutboxSourceKind {
         match self {
             Self::Http => "http",
             Self::Kv => "kv",
+            Self::Docs => "docs",
             Self::DeadLetter => "dead_letter",
+            Self::Cron => "cron",
+            Self::Files => "files",
+            Self::Pubsub => "pubsub",
         }
     }
 
@@ -40,7 +52,11 @@ impl OutboxSourceKind {
         match s {
             "http" => Some(Self::Http),
             "kv" => Some(Self::Kv),
+            "docs" => Some(Self::Docs),
             "dead_letter" => Some(Self::DeadLetter),
+            "cron" => Some(Self::Cron),
+            "files" => Some(Self::Files),
+            "pubsub" => Some(Self::Pubsub),
             _ => None,
         }
     }

crates/manager-core/src/pubsub_repo.rs

@@ -0,0 +1,118 @@
+//! `PubsubRepo` — publish-time fan-out for the v1.1.5 `pubsub::*` SDK.
+//!
+//! `publish_durable` writes one outbox row per matching enabled `pubsub`
+//! trigger, all inside a single transaction so a partial fan-out (some
+//! subscribers got rows, others didn't, then a crash) can't happen.
+//! Each delivery row then retries / dead-letters independently through
+//! the existing dispatcher — no pub/sub-specific dispatch branching.
+//!
+//! Topic pattern matching runs in Rust (`picloud_shared::topic_matches`)
+//! against the small set of the app's enabled pubsub triggers, keeping
+//! the SELECT trivial. v1.2 can add a topic-trie index if fan-out
+//! becomes a hot path.
+
+use async_trait::async_trait;
+use picloud_shared::{topic_matches, AdminUserId, AppId, ExecutionId};
+use sqlx::PgPool;
+use uuid::Uuid;
+
+#[derive(Debug, thiserror::Error)]
+pub enum PubsubRepoError {
+    #[error("database error: {0}")]
+    Db(#[from] sqlx::Error),
+}
+
+/// The execution-context bits a fan-out needs to stamp onto each outbox
+/// row. Derived from the publishing script's `SdkCallCx`.
+#[derive(Debug, Clone, Copy)]
+pub struct PublishCtx {
+    pub app_id: AppId,
+    pub origin_principal: Option<AdminUserId>,
+    pub trigger_depth: u32,
+    pub root_execution_id: ExecutionId,
+}
+
+#[async_trait]
+pub trait PubsubRepo: Send + Sync {
+    /// Fan out a publish to every matching enabled pubsub trigger in
+    /// `ctx.app_id`, inserting one outbox row each in a SINGLE
+    /// transaction. `event_payload` is the serialized
+    /// `TriggerEvent::Pubsub`. Returns the number of delivery rows
+    /// written (0 when no trigger matched — the publish still succeeds).
+    async fn fan_out_publish(
+        &self,
+        ctx: PublishCtx,
+        topic: &str,
+        event_payload: serde_json::Value,
+    ) -> Result<u32, PubsubRepoError>;
+}
+
+pub struct PostgresPubsubRepo {
+    pool: PgPool,
+}
+
+impl PostgresPubsubRepo {
+    #[must_use]
+    pub fn new(pool: PgPool) -> Self {
+        Self { pool }
+    }
+}
+
+#[derive(sqlx::FromRow)]
+struct PubsubTriggerRow {
+    id: Uuid,
+    script_id: Uuid,
+    topic_pattern: String,
+}
+
+#[async_trait]
+impl PubsubRepo for PostgresPubsubRepo {
+    async fn fan_out_publish(
+        &self,
+        ctx: PublishCtx,
+        topic: &str,
+        event_payload: serde_json::Value,
+    ) -> Result<u32, PubsubRepoError> {
+        let mut tx = self.pool.begin().await?;
+
+        // Load all enabled pubsub triggers for the app; filter by topic
+        // pattern in Rust (keeps the query simple, honours the
+        // empty/`*`/prefix semantics without teaching SQL about globs).
+        let rows: Vec<PubsubTriggerRow> = sqlx::query_as(
+            "SELECT t.id, t.script_id, d.topic_pattern \
+             FROM triggers t \
+             JOIN pubsub_trigger_details d ON d.trigger_id = t.id \
+             WHERE t.app_id = $1 AND t.kind = 'pubsub' AND t.enabled = TRUE",
+        )
+        .bind(ctx.app_id.into_inner())
+        .fetch_all(&mut *tx)
+        .await?;
+
+        let mut written: u32 = 0;
+        for r in rows {
+            if !topic_matches(&r.topic_pattern, topic) {
+                continue;
+            }
+            sqlx::query(
+                "INSERT INTO outbox ( \
+                    app_id, source_kind, trigger_id, script_id, reply_to, \
+                    payload, origin_principal, trigger_depth, root_execution_id \
+                 ) VALUES ($1, 'pubsub', $2, $3, NULL, $4, $5, $6, $7)",
+            )
+            .bind(ctx.app_id.into_inner())
+            .bind(r.id)
+            .bind(r.script_id)
+            .bind(&event_payload)
+            .bind(ctx.origin_principal.map(AdminUserId::into_inner))
+            .bind(i32::try_from(ctx.trigger_depth.saturating_add(1)).unwrap_or(1))
+            .bind(ctx.root_execution_id.into_inner())
+            .execute(&mut *tx)
+            .await?;
+            written += 1;
+        }
+
+        // Commit once — all rows or none.
+        tx.commit().await?;
+        Ok(written)
+    }
+}

crates/manager-core/src/pubsub_service.rs

@@ -0,0 +1,320 @@
+//! `PubsubServiceImpl` — wires `PubsubRepo` underneath the
+//! `picloud_shared::PubsubService` trait scripts see via the Rhai
+//! bridge.
+//!
+//! Mirrors the other stateful services: script-as-gate authz
+//! (`AppPubsubPublish`, skipped when `cx.principal` is `None`), with the
+//! backend doing a publish-time outbox fan-out instead of a row write.
+//! No `ServiceEventEmitter` here — pub/sub publishes directly to the
+//! outbox; it doesn't mutate local data that other triggers observe.
+
+use std::sync::Arc;
+
+use async_trait::async_trait;
+use picloud_shared::{PubsubError, PubsubService, SdkCallCx, TriggerEvent};
+
+use crate::authz::{self, AuthzRepo, Capability};
+use crate::pubsub_repo::{PublishCtx, PubsubRepo, PubsubRepoError};
+
+pub struct PubsubServiceImpl {
+    repo: Arc<dyn PubsubRepo>,
+    authz: Arc<dyn AuthzRepo>,
+}
+
+impl PubsubServiceImpl {
+    #[must_use]
+    pub fn new(repo: Arc<dyn PubsubRepo>, authz: Arc<dyn AuthzRepo>) -> Self {
+        Self { repo, authz }
+    }
+
+    async fn check_publish(&self, cx: &SdkCallCx) -> Result<(), PubsubError> {
+        if let Some(ref principal) = cx.principal {
+            authz::require(
+                &*self.authz,
+                principal,
+                Capability::AppPubsubPublish(cx.app_id),
+            )
+            .await
+            .map_err(|_| PubsubError::Forbidden)?;
+        }
+        Ok(())
+    }
+}
+
+impl From<PubsubRepoError> for PubsubError {
+    fn from(e: PubsubRepoError) -> Self {
+        Self::Unavailable(e.to_string())
+    }
+}
+
+#[async_trait]
+impl PubsubService for PubsubServiceImpl {
+    async fn publish_durable(
+        &self,
+        cx: &SdkCallCx,
+        topic: &str,
+        message: serde_json::Value,
+    ) -> Result<(), PubsubError> {
+        if topic.trim().is_empty() {
+            return Err(PubsubError::EmptyTopic);
+        }
+        self.check_publish(cx).await?;
+
+        // `published_at` is stamped on the manager side so every
+        // delivery agrees on one instant.
+        let event = TriggerEvent::Pubsub {
+            topic: topic.to_string(),
+            message,
+            published_at: chrono::Utc::now(),
+        };
+        let payload = serde_json::to_value(&event)
+            .map_err(|e| PubsubError::Rejected(format!("event serialize: {e}")))?;
+
+        let publish_ctx = PublishCtx {
+            app_id: cx.app_id,
+            origin_principal: cx.principal.as_ref().map(|p| p.user_id),
+            trigger_depth: cx.trigger_depth,
+            root_execution_id: cx.root_execution_id,
+        };
+        // No matching triggers → 0 rows written, publish still succeeds.
+        self.repo
+            .fan_out_publish(publish_ctx, topic, payload)
+            .await?;
+        Ok(())
+    }
+}
+
+// ----------------------------------------------------------------------------
+// Tests — in-memory PubsubRepo so unit tests don't need Postgres. The
+// real transactional fan-out is covered against a live DB by the
+// integration suite; the in-memory fake models the all-or-nothing
+// commit so the rollback semantics can be asserted without a DB.
+// ----------------------------------------------------------------------------
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::authz::{AuthzError, AuthzRepo};
+    use async_trait::async_trait;
+    use picloud_shared::{
+        topic_matches, AdminUserId, AppId, AppRole, ExecutionId, InstanceRole, Principal,
+        RequestId, ScriptId, UserId,
+    };
+    use std::sync::Mutex;
+
+    /// In-memory pubsub repo. Holds a set of `(app, pattern)`
+    /// subscriptions and records the outbox rows a publish would write.
+    /// `fail_at` simulates a mid-fan-out INSERT failure: when set to
+    /// `Some(n)`, the n-th (1-indexed) matching row errors and NOTHING
+    /// is recorded — modelling the single-transaction rollback.
+    struct InMemoryPubsubRepo {
+        subs: Vec<(AppId, String)>,
+        written: Mutex<Vec<(AppId, String)>>,
+        fail_at: Option<usize>,
+    }
+
+    impl InMemoryPubsubRepo {
+        fn new(subs: Vec<(AppId, String)>) -> Self {
+            Self {
+                subs,
+                written: Mutex::new(Vec::new()),
+                fail_at: None,
+            }
+        }
+        fn written_count(&self) -> usize {
+            self.written.lock().unwrap().len()
+        }
+    }
+
+    #[async_trait]
+    impl PubsubRepo for InMemoryPubsubRepo {
+        async fn fan_out_publish(
+            &self,
+            ctx: PublishCtx,
+            topic: &str,
+            _event_payload: serde_json::Value,
+        ) -> Result<u32, PubsubRepoError> {
+            let matches: Vec<&(AppId, String)> = self
+                .subs
+                .iter()
+                .filter(|(a, pat)| *a == ctx.app_id && topic_matches(pat, topic))
+                .collect();
+            let mut staged = Vec::new();
+            for (i, _) in matches.iter().enumerate() {
+                if self.fail_at == Some(i + 1) {
+                    // Rollback: nothing was committed.
+                    return Err(PubsubRepoError::Db(sqlx::Error::Protocol(
+                        "simulated insert failure".into(),
+                    )));
+                }
+                staged.push((ctx.app_id, topic.to_string()));
+            }
+            let n = staged.len();
+            self.written.lock().unwrap().extend(staged);
+            Ok(u32::try_from(n).unwrap_or(u32::MAX))
+        }
+    }
+
+    #[derive(Default)]
+    struct DenyingAuthzRepo;
+    #[async_trait]
+    impl AuthzRepo for DenyingAuthzRepo {
+        async fn membership(
+            &self,
+            _user_id: UserId,
+            _app_id: AppId,
+        ) -> Result<Option<AppRole>, AuthzError> {
+            Ok(None)
+        }
+    }
+
+    #[derive(Default)]
+    struct EditorAuthzRepo;
+    #[async_trait]
+    impl AuthzRepo for EditorAuthzRepo {
+        async fn membership(
+            &self,
+            _user_id: UserId,
+            _app_id: AppId,
+        ) -> Result<Option<AppRole>, AuthzError> {
+            Ok(Some(AppRole::Editor))
+        }
+    }
+
+    fn anon_cx(app_id: AppId) -> SdkCallCx {
+        SdkCallCx {
+            app_id,
+            script_id: ScriptId::new(),
+            principal: None,
+            execution_id: ExecutionId::new(),
+            request_id: RequestId::new(),
+            trigger_depth: 0,
+            root_execution_id: ExecutionId::new(),
+            is_dead_letter_handler: false,
+            event: None,
+        }
+    }
+
+    fn member_cx(app_id: AppId) -> SdkCallCx {
+        SdkCallCx {
+            principal: Some(Principal {
+                user_id: AdminUserId::new(),
+                instance_role: InstanceRole::Member,
+                scopes: None,
+                app_binding: None,
+            }),
+            ..anon_cx(app_id)
+        }
+    }
+
+    fn svc(repo: Arc<dyn PubsubRepo>, authz: Arc<dyn AuthzRepo>) -> PubsubServiceImpl {
+        PubsubServiceImpl::new(repo, authz)
+    }
+
+    #[tokio::test]
+    async fn publish_writes_one_row_per_matching_trigger() {
+        let app = AppId::new();
+        let repo = Arc::new(InMemoryPubsubRepo::new(vec![
+            (app, "user.*".into()),
+            (app, "user.created".into()),
+            (app, "order.*".into()), // does not match
+        ]));
+        let files = svc(repo.clone(), Arc::new(DenyingAuthzRepo));
+        files
+            .publish_durable(&anon_cx(app), "user.created", serde_json::json!({"id": 1}))
+            .await
+            .unwrap();
+        // Two of the three subscriptions match "user.created".
+        assert_eq!(repo.written_count(), 2);
+    }
+
+    #[tokio::test]
+    async fn no_matching_trigger_succeeds_silently() {
+        let app = AppId::new();
+        let repo = Arc::new(InMemoryPubsubRepo::new(vec![(app, "order.*".into())]));
+        let svc = svc(repo.clone(), Arc::new(DenyingAuthzRepo));
+        svc.publish_durable(&anon_cx(app), "user.created", serde_json::json!(1))
+            .await
+            .unwrap();
+        assert_eq!(repo.written_count(), 0);
+    }
+
+    #[tokio::test]
+    async fn empty_topic_rejected() {
+        let app = AppId::new();
+        let repo = Arc::new(InMemoryPubsubRepo::new(vec![]));
+        let svc = svc(repo, Arc::new(DenyingAuthzRepo));
+        let err = svc
+            .publish_durable(&anon_cx(app), "   ", serde_json::json!(1))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, PubsubError::EmptyTopic));
+    }
+
+    #[tokio::test]
+    async fn cross_app_isolation() {
+        let app_a = AppId::new();
+        let app_b = AppId::new();
+        // The only subscription belongs to app B.
+        let repo = Arc::new(InMemoryPubsubRepo::new(vec![(app_b, "*".into())]));
+        let svc = svc(repo.clone(), Arc::new(DenyingAuthzRepo));
+        // App A publishes — app B's trigger must NOT fire.
+        svc.publish_durable(&anon_cx(app_a), "user.created", serde_json::json!(1))
+            .await
+            .unwrap();
+        assert_eq!(repo.written_count(), 0);
+    }
+
+    #[tokio::test]
+    async fn fan_out_is_transactional_all_or_nothing() {
+        let app = AppId::new();
+        let mut repo = InMemoryPubsubRepo::new(vec![
+            (app, "*".into()),
+            (app, "user.*".into()),
+            (app, "user.created".into()),
+        ]);
+        repo.fail_at = Some(3); // fail on the 3rd matching insert
+        let repo = Arc::new(repo);
+        let svc = svc(repo.clone(), Arc::new(DenyingAuthzRepo));
+        let err = svc
+            .publish_durable(&anon_cx(app), "user.created", serde_json::json!(1))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, PubsubError::Unavailable(_)));
+        // Rollback: no partial fan-out survived.
+        assert_eq!(repo.written_count(), 0);
+    }
+
+    #[tokio::test]
+    async fn anonymous_cx_skips_authz() {
+        let app = AppId::new();
+        let repo = Arc::new(InMemoryPubsubRepo::new(vec![]));
+        let svc = svc(repo, Arc::new(DenyingAuthzRepo));
+        // No principal → no authz check even with a denying repo.
+        svc.publish_durable(&anon_cx(app), "t", serde_json::json!(1))
+            .await
+            .unwrap();
+    }
+
+    #[tokio::test]
+    async fn member_without_role_is_forbidden() {
+        let app = AppId::new();
+        let repo = Arc::new(InMemoryPubsubRepo::new(vec![]));
+        let svc = svc(repo, Arc::new(DenyingAuthzRepo));
+        let err = svc
+            .publish_durable(&member_cx(app), "t", serde_json::json!(1))
+            .await
+            .unwrap_err();
+        assert!(matches!(err, PubsubError::Forbidden));
+    }
+
+    #[tokio::test]
+    async fn member_with_editor_role_allowed() {
+        let app = AppId::new();
+        let repo = Arc::new(InMemoryPubsubRepo::new(vec![]));
+        let svc = svc(repo, Arc::new(EditorAuthzRepo));
+        svc.publish_durable(&member_cx(app), "t", serde_json::json!(1))
+            .await
+            .unwrap();
+    }
+}

crates/manager-core/src/repo.rs

@@ -3,7 +3,8 @@ use std::collections::BTreeMap;
 use async_trait::async_trait;
 use picloud_orchestrator_core::{ResolverError, ScriptResolver};
 use picloud_shared::{
-    AdminUserId, AppId, ExecutionLog, ExecutionStatus, RequestId, Script, ScriptId, ScriptSandbox,
+    AdminUserId, AppId, ExecutionLog, ExecutionStatus, RequestId, Script, ScriptId, ScriptKind,
+    ScriptSandbox,
 };
 use sqlx::PgPool;
 
@@ -42,6 +43,27 @@ pub trait ScriptRepository: Send + Sync {
         patch: ScriptPatch,
     ) -> Result<Script, ScriptRepositoryError>;
     async fn delete(&self, id: ScriptId) -> Result<(), ScriptRepositoryError>;
+
+    /// v1.1.3: how many routes reference this script. Used by the
+    /// API layer to refuse `endpoint → module` kind changes when the
+    /// script is still bound to user-facing entry points.
+    async fn count_routes_for_script(
+        &self,
+        script_id: ScriptId,
+    ) -> Result<i64, ScriptRepositoryError>;
+
+    /// v1.1.3: how many triggers (kv / docs / dead-letter) target
+    /// this script. Same purpose as `count_routes_for_script`.
+    async fn count_triggers_for_script(
+        &self,
+        script_id: ScriptId,
+    ) -> Result<i64, ScriptRepositoryError>;
+
+    /// v1.1.3: list module dependencies of this script — the rows in
+    /// `script_imports` where `importer_script_id = script_id`. Used
+    /// by tests and (eventually) a dashboard "Imports" panel.
+    async fn list_imports(&self, script_id: ScriptId)
+        -> Result<Vec<Script>, ScriptRepositoryError>;
 }
 
 /// Inbound shape for create. Defaults match the migration's CHECK
@@ -52,11 +74,19 @@ pub struct NewScript {
     pub name: String,
     pub description: Option<String>,
     pub source: String,
+    /// Defaults to `Endpoint` if absent. `Module` scripts cannot be
+    /// bound to routes or used as trigger targets.
+    pub kind: ScriptKind,
     pub timeout_seconds: Option<i32>,
     pub memory_limit_mb: Option<i32>,
     /// Sandbox overrides; `None` means store an empty object (use
     /// platform defaults at exec time).
     pub sandbox: Option<ScriptSandbox>,
+    /// v1.1.3: literal-path `import "<name>"` declarations extracted
+    /// from the source. The repo writes these into `script_imports`
+    /// transactionally with the script row. Empty when validation
+    /// found no imports (the common case for endpoints today).
+    pub imports: Vec<String>,
 }
 
 /// Inbound shape for update. `None` fields are left untouched.
@@ -70,6 +100,15 @@ pub struct ScriptPatch {
     /// `Some(sandbox)` replaces the stored overrides wholesale (including
     /// `Some(empty)` to clear them); `None` leaves them untouched.
     pub sandbox: Option<ScriptSandbox>,
+    /// `Some(new_kind)` changes the script's role; the API layer
+    /// rejects unsafe transitions (e.g. endpoint→module when routes
+    /// or triggers reference the script).
+    pub kind: Option<ScriptKind>,
+    /// v1.1.3: when `source` is also `Some`, the repo replaces the
+    /// `script_imports` edges for this script with these names.
+    /// `None` keeps the existing edges untouched (a name/description
+    /// edit alone shouldn't touch the dep graph).
+    pub imports: Option<Vec<String>>,
 }
 
 pub struct PostgresScriptRepository {
@@ -88,14 +127,18 @@ impl PostgresScriptRepository {
     }
 }
 
+/// Columns selected from `scripts` everywhere — kept in one constant so
+/// adding `kind` (v1.1.3) and future columns can't accidentally skip
+/// one query.
+const SCRIPT_SELECT_COLS: &str = "id, app_id, name, description, version, source, kind, \
+                                  timeout_seconds, memory_limit_mb, sandbox, created_at, updated_at";
+
 #[async_trait]
 impl ScriptRepository for PostgresScriptRepository {
     async fn get(&self, id: ScriptId) -> Result<Option<Script>, ScriptRepositoryError> {
-        let row = sqlx::query_as::<_, ScriptRow>(
-            "SELECT id, app_id, name, description, version, source, \
-                    timeout_seconds, memory_limit_mb, sandbox, created_at, updated_at \
-             FROM scripts WHERE id = $1",
-        )
+        let row = sqlx::query_as::<_, ScriptRow>(&format!(
+            "SELECT {SCRIPT_SELECT_COLS} FROM scripts WHERE id = $1"
+        ))
         .bind(id.into_inner())
         .fetch_optional(&self.pool)
         .await?;
@@ -103,22 +146,18 @@ impl ScriptRepository for PostgresScriptRepository {
     }
 
     async fn list(&self) -> Result<Vec<Script>, ScriptRepositoryError> {
-        let rows = sqlx::query_as::<_, ScriptRow>(
-            "SELECT id, app_id, name, description, version, source, \
-                    timeout_seconds, memory_limit_mb, sandbox, created_at, updated_at \
-             FROM scripts ORDER BY name",
-        )
+        let rows = sqlx::query_as::<_, ScriptRow>(&format!(
+            "SELECT {SCRIPT_SELECT_COLS} FROM scripts ORDER BY name"
+        ))
         .fetch_all(&self.pool)
         .await?;
         Ok(rows.into_iter().map(Into::into).collect())
     }
 
     async fn list_for_app(&self, app_id: AppId) -> Result<Vec<Script>, ScriptRepositoryError> {
-        let rows = sqlx::query_as::<_, ScriptRow>(
-            "SELECT id, app_id, name, description, version, source, \
-                    timeout_seconds, memory_limit_mb, sandbox, created_at, updated_at \
-             FROM scripts WHERE app_id = $1 ORDER BY name",
-        )
+        let rows = sqlx::query_as::<_, ScriptRow>(&format!(
+            "SELECT {SCRIPT_SELECT_COLS} FROM scripts WHERE app_id = $1 ORDER BY name"
+        ))
         .bind(app_id.into_inner())
         .fetch_all(&self.pool)
         .await?;
@@ -129,14 +168,17 @@ impl ScriptRepository for PostgresScriptRepository {
         &self,
         user_id: AdminUserId,
     ) -> Result<Vec<Script>, ScriptRepositoryError> {
-        let rows = sqlx::query_as::<_, ScriptRow>(
-            "SELECT s.id, s.app_id, s.name, s.description, s.version, s.source, \
-                    s.timeout_seconds, s.memory_limit_mb, s.sandbox, s.created_at, s.updated_at \
-             FROM scripts s \
+        let cols = SCRIPT_SELECT_COLS
+            .split(", ")
+            .map(|c| format!("s.{c}"))
+            .collect::<Vec<_>>()
+            .join(", ");
+        let rows = sqlx::query_as::<_, ScriptRow>(&format!(
+            "SELECT {cols} FROM scripts s \
              JOIN app_members m ON m.app_id = s.app_id \
              WHERE m.user_id = $1 \
-             ORDER BY s.name",
-        )
+             ORDER BY s.name"
+        ))
         .bind(user_id.into_inner())
         .fetch_all(&self.pool)
         .await?;
@@ -146,34 +188,42 @@ impl ScriptRepository for PostgresScriptRepository {
     async fn create(&self, input: NewScript) -> Result<Script, ScriptRepositoryError> {
         let sandbox_json = serde_json::to_value(input.sandbox.unwrap_or_default())
             .unwrap_or_else(|_| serde_json::json!({}));
-        let res = sqlx::query_as::<_, ScriptRow>(
+        let mut tx = self.pool.begin().await?;
+        let res = sqlx::query_as::<_, ScriptRow>(&format!(
             "INSERT INTO scripts ( \
-                app_id, name, description, source, \
+                app_id, name, description, source, kind, \
                 timeout_seconds, memory_limit_mb, sandbox \
-             ) VALUES ($1, $2, $3, $4, COALESCE($5, 30), COALESCE($6, 256), $7) \
-             RETURNING id, app_id, name, description, version, source, \
-                       timeout_seconds, memory_limit_mb, sandbox, created_at, updated_at",
-        )
+             ) VALUES ($1, $2, $3, $4, $5, COALESCE($6, 30), COALESCE($7, 256), $8) \
+             RETURNING {SCRIPT_SELECT_COLS}"
+        ))
         .bind(input.app_id.into_inner())
         .bind(&input.name)
         .bind(input.description.as_deref())
         .bind(&input.source)
+        .bind(input.kind.as_str())
         .bind(input.timeout_seconds)
         .bind(input.memory_limit_mb)
         .bind(sandbox_json)
-        .fetch_one(&self.pool)
+        .fetch_one(&mut *tx)
         .await;
 
-        match res {
-            Ok(row) => Ok(row.into()),
+        let script: Script = match res {
+            Ok(row) => row.into(),
             Err(sqlx::Error::Database(e)) if e.is_unique_violation() => {
-                Err(ScriptRepositoryError::Conflict(format!(
+                return Err(ScriptRepositoryError::Conflict(format!(
                     "a script named {:?} already exists in this app",
                     input.name
-                )))
-            }
-            Err(e) => Err(e.into()),
+                )));
             }
+            Err(e) => return Err(e.into()),
+        };
+
+        // Dep-graph: write any literal-path imports declared in the
+        // source. Unresolved names (the referenced module doesn't
+        // exist yet) are silently skipped — best-effort.
+        replace_imports_tx(&mut tx, script.id, script.app_id, &input.imports).await?;
+        tx.commit().await?;
+        Ok(script)
     }
 
     async fn update(
@@ -192,7 +242,8 @@ impl ScriptRepository for PostgresScriptRepository {
             .sandbox
             .as_ref()
             .map(|s| serde_json::to_value(s).unwrap_or_else(|_| serde_json::json!({})));
-        let res = sqlx::query_as::<_, ScriptRow>(
+        let mut tx = self.pool.begin().await?;
+        let res = sqlx::query_as::<_, ScriptRow>(&format!(
             "UPDATE scripts SET \
                 name = COALESCE($2, name), \
                 description = CASE WHEN $3::bool THEN $4 ELSE description END, \
@@ -200,12 +251,12 @@ impl ScriptRepository for PostgresScriptRepository {
                 timeout_seconds = COALESCE($6, timeout_seconds), \
                 memory_limit_mb = COALESCE($7, memory_limit_mb), \
                 sandbox = COALESCE($8, sandbox), \
+                kind = COALESCE($9, kind), \
                 version = version + 1, \
                 updated_at = NOW() \
              WHERE id = $1 \
-             RETURNING id, app_id, name, description, version, source, \
-                       timeout_seconds, memory_limit_mb, sandbox, created_at, updated_at",
-        )
+             RETURNING {SCRIPT_SELECT_COLS}"
+        ))
         .bind(id.into_inner())
         .bind(patch.name.as_deref())
         .bind(patch.description.is_some())
@@ -214,19 +265,30 @@ impl ScriptRepository for PostgresScriptRepository {
         .bind(patch.timeout_seconds)
         .bind(patch.memory_limit_mb)
         .bind(sandbox_json)
-        .fetch_optional(&self.pool)
+        .bind(patch.kind.map(ScriptKind::as_str))
+        .fetch_optional(&mut *tx)
         .await;
 
-        match res {
-            Ok(Some(row)) => Ok(row.into()),
-            Ok(None) => Err(ScriptRepositoryError::NotFound(id)),
+        let script: Script = match res {
+            Ok(Some(row)) => row.into(),
+            Ok(None) => return Err(ScriptRepositoryError::NotFound(id)),
             Err(sqlx::Error::Database(e)) if e.is_unique_violation() => {
-                Err(ScriptRepositoryError::Conflict(
+                return Err(ScriptRepositoryError::Conflict(
                     "a script with that name already exists in this app".into(),
-                ))
+                ));
             }
-            Err(e) => Err(e.into()),
+            Err(e) => return Err(e.into()),
+        };
+
+        // Replace imports only when the caller has a fresh list (i.e.
+        // the source actually changed and the validator re-extracted
+        // imports). A name-only or description-only edit leaves the
+        // dep graph alone.
+        if let Some(imports) = patch.imports.as_deref() {
+            replace_imports_tx(&mut tx, script.id, script.app_id, imports).await?;
         }
+        tx.commit().await?;
+        Ok(script)
     }
 
     async fn delete(&self, id: ScriptId) -> Result<(), ScriptRepositoryError> {
@@ -239,6 +301,85 @@ impl ScriptRepository for PostgresScriptRepository {
         }
         Ok(())
     }
+
+    async fn count_routes_for_script(
+        &self,
+        script_id: ScriptId,
+    ) -> Result<i64, ScriptRepositoryError> {
+        let n: (i64,) = sqlx::query_as("SELECT COUNT(*) FROM routes WHERE script_id = $1")
+            .bind(script_id.into_inner())
+            .fetch_one(&self.pool)
+            .await?;
+        Ok(n.0)
+    }
+
+    async fn count_triggers_for_script(
+        &self,
+        script_id: ScriptId,
+    ) -> Result<i64, ScriptRepositoryError> {
+        let n: (i64,) = sqlx::query_as("SELECT COUNT(*) FROM triggers WHERE script_id = $1")
+            .bind(script_id.into_inner())
+            .fetch_one(&self.pool)
+            .await?;
+        Ok(n.0)
+    }
+
+    async fn list_imports(
+        &self,
+        script_id: ScriptId,
+    ) -> Result<Vec<Script>, ScriptRepositoryError> {
+        let cols = SCRIPT_SELECT_COLS
+            .split(", ")
+            .map(|c| format!("s.{c}"))
+            .collect::<Vec<_>>()
+            .join(", ");
+        let rows = sqlx::query_as::<_, ScriptRow>(&format!(
+            "SELECT {cols} FROM scripts s \
+             JOIN script_imports i ON i.imported_script_id = s.id \
+             WHERE i.importer_script_id = $1 \
+             ORDER BY s.name"
+        ))
+        .bind(script_id.into_inner())
+        .fetch_all(&self.pool)
+        .await?;
+        Ok(rows.into_iter().map(Into::into).collect())
+    }
+}
+
+/// Replace the `script_imports` edges for `importer` with rows derived
+/// from `import_names`. Names that don't resolve to a `kind = 'module'`
+/// script in the same app are silently skipped (best-effort dep graph).
+async fn replace_imports_tx(
+    tx: &mut sqlx::Transaction<'_, sqlx::Postgres>,
+    importer: ScriptId,
+    app_id: AppId,
+    import_names: &[String],
+) -> Result<(), ScriptRepositoryError> {
+    sqlx::query("DELETE FROM script_imports WHERE importer_script_id = $1")
+        .bind(importer.into_inner())
+        .execute(&mut **tx)
+        .await?;
+    if import_names.is_empty() {
+        return Ok(());
+    }
+    // Insert with ON CONFLICT DO NOTHING in case the source declares
+    // `import "x"` twice — the dep graph stores each pair at most once.
+    sqlx::query(
+        "INSERT INTO script_imports (app_id, importer_script_id, imported_script_id) \
+         SELECT $1, $2, s.id \
+         FROM scripts s \
+         WHERE s.app_id = $1 \
+           AND s.kind = 'module' \
+           AND s.id <> $2 \
+           AND s.name = ANY($3) \
+         ON CONFLICT DO NOTHING",
+    )
+    .bind(app_id.into_inner())
+    .bind(importer.into_inner())
+    .bind(import_names)
+    .execute(&mut **tx)
+    .await?;
+    Ok(())
 }
 
 /// Row shape mirroring the `scripts` table for sqlx FromRow.
@@ -250,6 +391,10 @@ struct ScriptRow {
     description: Option<String>,
     version: i32,
     source: String,
+    /// v1.1.3: 'endpoint' | 'module'. Stored as TEXT with a CHECK
+    /// constraint so we don't need a Postgres enum (avoiding the
+    /// migration churn of adding values later).
+    kind: String,
     timeout_seconds: i32,
     memory_limit_mb: i32,
     sandbox: serde_json::Value,
@@ -264,6 +409,10 @@ impl From<ScriptRow> for Script {
         // fall back to an empty ScriptSandbox rather than poisoning a
         // list response.
         let sandbox = serde_json::from_value(r.sandbox).unwrap_or_default();
+        // Defensive: if a row's `kind` somehow falls outside the CHECK
+        // constraint, treat it as Endpoint (the safe default — won't
+        // grant a row import-target status it doesn't have).
+        let kind = ScriptKind::parse_str(&r.kind).unwrap_or(ScriptKind::Endpoint);
         Self {
             id: r.id.into(),
             app_id: r.app_id.into(),
@@ -271,6 +420,7 @@ impl From<ScriptRow> for Script {
             description: r.description,
             version: r.version,
             source: r.source,
+            kind,
             timeout_seconds: u32::try_from(r.timeout_seconds).unwrap_or(30),
             memory_limit_mb: u32::try_from(r.memory_limit_mb).unwrap_or(256),
             sandbox,

crates/manager-core/src/route_admin.rs

@@ -184,6 +184,17 @@ async fn create_route<RR: RouteRepository, SR: ScriptRepository>(
     )
     .await?;
 
+    // v1.1.3: module scripts have no executable entry point — they're
+    // libraries imported by other scripts. Reject route bindings here
+    // before we touch the routes table.
+    if script.kind == picloud_shared::ScriptKind::Module {
+        return Err(RouteApiError::BadRequest(format!(
+            "script {script_id} has kind=module; modules are imported, \
+             not bound to routes — switch the script to kind=endpoint \
+             or attach this route to a different script"
+        )));
+    }
+
     // Validate the route's host is consistent with one of the app's
     // domain claims. `HostKind::Any` is always permitted (catches every
     // host the app already owns). Specific hosts must match a claim.

crates/manager-core/src/ssrf.rs

@@ -0,0 +1,457 @@
+//! SSRF deny-list — the load-bearing security mechanism behind the
+//! v1.1.4 `http::*` SDK.
+//!
+//! The policy is applied to the **resolved IP address**, not the
+//! hostname. That is the DNS-rebinding defense: a hostname that
+//! resolves to a public IP at lookup time and a private IP at connect
+//! time is not exploitable, because reqwest re-runs every connection
+//! (including post-redirect hops) through [`SsrfResolver`], which
+//! filters the address list before the socket is opened.
+//!
+//! [`SsrfPolicy::check`] returns a CIDR-*category* reason on denial
+//! (e.g. `"loopback"`, `"private"`) — never the IP itself, so the
+//! script-visible error can't be used to map the internal network.
+//!
+//! `PICLOUD_HTTP_ALLOW_PRIVATE=true` flips `allow_private`, which
+//! short-circuits every check to allow. That is dev/test-only and the
+//! binary logs a startup warning when it's set.
+
+use std::future::Future;
+use std::net::{IpAddr, Ipv4Addr, Ipv6Addr, SocketAddr};
+use std::pin::Pin;
+use std::sync::Arc;
+
+use reqwest::dns::{Addrs, Name, Resolve, Resolving};
+
+/// Decision policy for a single resolved IP. Cheap to clone (one bool).
+#[derive(Debug, Clone, Copy)]
+pub struct SsrfPolicy {
+    /// When true, every address is allowed — the entire deny-list is
+    /// disabled. Set from `PICLOUD_HTTP_ALLOW_PRIVATE`. Dev/test only.
+    pub allow_private: bool,
+}
+
+impl SsrfPolicy {
+    #[must_use]
+    pub const fn new(allow_private: bool) -> Self {
+        Self { allow_private }
+    }
+
+    /// `Ok(())` if the IP may be connected to; `Err(reason)` with a
+    /// CIDR-category label otherwise. The reason is safe to surface to
+    /// a script — it never contains the address.
+    ///
+    /// # Errors
+    ///
+    /// Returns the deny reason when `ip` falls in a blocked range and
+    /// `allow_private` is false.
+    pub fn check(&self, ip: IpAddr) -> Result<(), &'static str> {
+        if self.allow_private {
+            return Ok(());
+        }
+        match ip {
+            IpAddr::V4(v4) => check_v4(v4),
+            IpAddr::V6(v6) => check_v6(v6),
+        }
+    }
+
+    #[must_use]
+    pub fn is_allowed(&self, ip: IpAddr) -> bool {
+        self.check(ip).is_ok()
+    }
+}
+
+/// IPv4 deny-list. Order doesn't matter (ranges are disjoint by
+/// construction); first match wins for the reason label.
+// Several arms share a reason ("private") for distinct CIDRs — keeping
+// them separate documents each blocked range explicitly.
+#[allow(clippy::match_same_arms)]
+fn check_v4(ip: Ipv4Addr) -> Result<(), &'static str> {
+    let o = ip.octets();
+    match o {
+        [127, ..] => Err("loopback"),
+        [0, ..] => Err("unspecified"), // 0.0.0.0/8 "this network"
+        [10, ..] => Err("private"),
+        [172, b, ..] if (16..=31).contains(&b) => Err("private"),
+        [192, 168, ..] => Err("private"),
+        [169, 254, ..] => Err("link-local"), // includes cloud metadata 169.254.169.254
+        [100, b, ..] if (64..=127).contains(&b) => Err("carrier-grade-nat"),
+        [224..=239, ..] => Err("multicast"),
+        [240..=255, ..] => Err("reserved"),
+        _ => Ok(()),
+    }
+}
+
+/// IPv6 deny-list. IPv4-mapped addresses (`::ffff:0:0/96`) re-run the
+/// v4 deny-list against the embedded address.
+fn check_v6(ip: Ipv6Addr) -> Result<(), &'static str> {
+    // IPv4-mapped (::ffff:a.b.c.d) — re-check the embedded v4 address
+    // so a mapped private/loopback address can't sneak through.
+    if let Some(v4) = ip.to_ipv4_mapped() {
+        return check_v4(v4);
+    }
+    if ip == Ipv6Addr::LOCALHOST {
+        return Err("loopback");
+    }
+    if ip == Ipv6Addr::UNSPECIFIED {
+        return Err("unspecified");
+    }
+    let seg0 = ip.segments()[0];
+    if seg0 & 0xffc0 == 0xfe80 {
+        return Err("link-local"); // fe80::/10
+    }
+    if seg0 & 0xfe00 == 0xfc00 {
+        return Err("unique-local"); // fc00::/7
+    }
+    if seg0 & 0xff00 == 0xff00 {
+        return Err("multicast"); // ff00::/8
+    }
+    Ok(())
+}
+
+/// Marker error returned by the resolver when *every* resolved address
+/// for a host was denied. reqwest wraps this into a connect error; the
+/// `http_service` impl walks the source chain for the
+/// `"blocked by SSRF policy:"` prefix to surface a clean
+/// [`crate::http_service::HttpError::Ssrf`] instead of a generic DNS
+/// failure. Keeping the reason a category label means no IP leaks.
+#[derive(Debug)]
+struct SsrfBlocked {
+    reason: &'static str,
+}
+
+impl std::fmt::Display for SsrfBlocked {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        write!(f, "blocked by SSRF policy: {}", self.reason)
+    }
+}
+
+impl std::error::Error for SsrfBlocked {}
+
+/// Prefix the resolver embeds in its error and the impl scans for.
+pub const SSRF_BLOCK_PREFIX: &str = "blocked by SSRF policy: ";
+
+/// Pluggable host→addresses lookup. Production uses the system
+/// resolver; tests inject a closure (e.g. to simulate DNS rebinding —
+/// a different address on a later call).
+pub type LookupFn = Arc<
+    dyn Fn(String) -> Pin<Box<dyn Future<Output = std::io::Result<Vec<SocketAddr>>> + Send>>
+        + Send
+        + Sync,
+>;
+
+fn system_lookup(
+    host: String,
+) -> Pin<Box<dyn Future<Output = std::io::Result<Vec<SocketAddr>>> + Send>> {
+    Box::pin(async move {
+        // Port 0 — reqwest overrides it with the real target port.
+        Ok(tokio::net::lookup_host((host.as_str(), 0u16))
+            .await?
+            .collect())
+    })
+}
+
+/// reqwest DNS resolver that delegates to the system resolver, then
+/// filters the address list through [`SsrfPolicy`]. Plugged in via
+/// `ClientBuilder::dns_resolver`, so it runs at the actual connection
+/// point — including for every redirect hop. This is the DNS-rebinding
+/// defense: filtering happens at connect time, not at URL-parse time.
+#[derive(Clone)]
+pub struct SsrfResolver {
+    policy: SsrfPolicy,
+    lookup: LookupFn,
+}
+
+impl SsrfResolver {
+    #[must_use]
+    pub fn new(policy: SsrfPolicy) -> Self {
+        Self {
+            policy,
+            lookup: Arc::new(system_lookup),
+        }
+    }
+
+    /// Construct with an injected lookup (tests only).
+    #[must_use]
+    pub fn with_lookup(policy: SsrfPolicy, lookup: LookupFn) -> Self {
+        Self { policy, lookup }
+    }
+}
+
+impl Resolve for SsrfResolver {
+    fn resolve(&self, name: Name) -> Resolving {
+        let policy = self.policy;
+        let lookup = self.lookup.clone();
+        let host = name.as_str().to_string();
+        Box::pin(async move {
+            let resolved: Vec<SocketAddr> = lookup(host)
+                .await
+                .map_err(|e| -> Box<dyn std::error::Error + Send + Sync> { Box::new(e) })?;
+
+            // Empty resolution → genuine DNS miss; let reqwest surface
+            // it as a normal "no addresses" error.
+            if resolved.is_empty() {
+                let addrs: Addrs = Box::new(std::iter::empty());
+                return Ok(addrs);
+            }
+
+            let mut allowed: Vec<SocketAddr> = Vec::with_capacity(resolved.len());
+            let mut last_reason: &'static str = "denied";
+            for sa in resolved {
+                match policy.check(sa.ip()) {
+                    Ok(()) => allowed.push(sa),
+                    Err(reason) => last_reason = reason,
+                }
+            }
+
+            // Resolution returned addresses but the policy denied them
+            // all → fail with the SSRF marker so the impl can report a
+            // policy block (not a generic DNS error).
+            if allowed.is_empty() {
+                let err: Box<dyn std::error::Error + Send + Sync> = Box::new(SsrfBlocked {
+                    reason: last_reason,
+                });
+                return Err(err);
+            }
+
+            let addrs: Addrs = Box::new(allowed.into_iter());
+            Ok(addrs)
+        })
+    }
+}
+
+/// Build the resolver. reqwest's `dns_resolver` is generic over a
+/// concrete `R: Resolve` (it stores `Arc<R>`), so this returns the
+/// concrete `Arc<SsrfResolver>` rather than a trait object.
+#[must_use]
+pub fn resolver(policy: SsrfPolicy) -> Arc<SsrfResolver> {
+    Arc::new(SsrfResolver::new(policy))
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use std::str::FromStr;
+
+    fn denied(ip: &str) -> &'static str {
+        let policy = SsrfPolicy::new(false);
+        policy
+            .check(IpAddr::from_str(ip).unwrap())
+            .expect_err(&format!("{ip} should be denied"))
+    }
+
+    fn allowed(ip: &str) {
+        let policy = SsrfPolicy::new(false);
+        policy
+            .check(IpAddr::from_str(ip).unwrap())
+            .unwrap_or_else(|r| panic!("{ip} should be allowed, denied as {r}"));
+    }
+
+    #[test]
+    fn denies_ipv4_loopback() {
+        assert_eq!(denied("127.0.0.1"), "loopback");
+        assert_eq!(denied("127.1.2.3"), "loopback");
+    }
+
+    #[test]
+    fn denies_ipv4_unspecified() {
+        assert_eq!(denied("0.0.0.0"), "unspecified");
+    }
+
+    #[test]
+    fn denies_rfc1918_private() {
+        assert_eq!(denied("10.0.0.1"), "private");
+        assert_eq!(denied("10.255.255.255"), "private");
+        assert_eq!(denied("172.16.0.1"), "private");
+        assert_eq!(denied("172.31.255.255"), "private");
+        assert_eq!(denied("192.168.0.1"), "private");
+    }
+
+    #[test]
+    fn allows_172_outside_private_range() {
+        // 172.15.x and 172.32.x are public — only 172.16.0.0/12 is private.
+        allowed("172.15.0.1");
+        allowed("172.32.0.1");
+    }
+
+    #[test]
+    fn denies_link_local_and_cloud_metadata() {
+        assert_eq!(denied("169.254.0.1"), "link-local");
+        // The cloud metadata endpoint is the canonical SSRF target.
+        assert_eq!(denied("169.254.169.254"), "link-local");
+    }
+
+    #[test]
+    fn denies_carrier_grade_nat() {
+        assert_eq!(denied("100.64.0.1"), "carrier-grade-nat");
+        assert_eq!(denied("100.127.255.255"), "carrier-grade-nat");
+        // 100.63.x and 100.128.x are outside 100.64.0.0/10.
+        allowed("100.63.0.1");
+        allowed("100.128.0.1");
+    }
+
+    #[test]
+    fn denies_multicast_and_reserved() {
+        assert_eq!(denied("224.0.0.1"), "multicast");
+        assert_eq!(denied("239.255.255.255"), "multicast");
+        assert_eq!(denied("240.0.0.1"), "reserved");
+        assert_eq!(denied("255.255.255.255"), "reserved");
+    }
+
+    #[test]
+    fn allows_public_ipv4() {
+        allowed("1.1.1.1");
+        allowed("8.8.8.8");
+        allowed("93.184.216.34"); // example.com
+    }
+
+    #[test]
+    fn denies_ipv6_loopback() {
+        assert_eq!(denied("::1"), "loopback");
+    }
+
+    #[test]
+    fn denies_ipv6_unspecified() {
+        assert_eq!(denied("::"), "unspecified");
+    }
+
+    #[test]
+    fn denies_ipv6_link_local() {
+        assert_eq!(denied("fe80::1"), "link-local");
+        assert_eq!(denied("febf:ffff::1"), "link-local");
+    }
+
+    #[test]
+    fn denies_ipv6_unique_local() {
+        assert_eq!(denied("fc00::1"), "unique-local");
+        assert_eq!(denied("fd12:3456::1"), "unique-local");
+    }
+
+    #[test]
+    fn denies_ipv6_multicast() {
+        assert_eq!(denied("ff00::1"), "multicast");
+        assert_eq!(denied("ff02::1"), "multicast");
+    }
+
+    #[test]
+    fn allows_public_ipv6() {
+        allowed("2606:4700:4700::1111"); // cloudflare
+        allowed("2001:4860:4860::8888"); // google
+    }
+
+    #[test]
+    fn ipv4_mapped_ipv6_rechecks_embedded_address() {
+        // ::ffff:127.0.0.1 must be denied via the embedded-v4 re-check.
+        assert_eq!(denied("::ffff:127.0.0.1"), "loopback");
+        assert_eq!(denied("::ffff:10.0.0.1"), "private");
+        assert_eq!(denied("::ffff:169.254.169.254"), "link-local");
+        // A mapped *public* address stays allowed.
+        allowed("::ffff:1.1.1.1");
+    }
+
+    #[test]
+    fn allow_private_disables_all_denials() {
+        let policy = SsrfPolicy::new(true);
+        for ip in ["127.0.0.1", "10.0.0.1", "169.254.169.254", "::1", "fe80::1"] {
+            assert!(policy.is_allowed(IpAddr::from_str(ip).unwrap()));
+        }
+    }
+
+    // --- resolver-path tests (the connect-time filter) ---
+
+    use std::sync::atomic::{AtomicUsize, Ordering};
+
+    fn name(s: &str) -> Name {
+        Name::from_str(s).unwrap()
+    }
+
+    fn fixed_lookup(addrs: Vec<SocketAddr>) -> LookupFn {
+        Arc::new(move |_host| {
+            let addrs = addrs.clone();
+            Box::pin(async move { Ok(addrs) })
+        })
+    }
+
+    #[tokio::test]
+    async fn resolver_returns_only_allowed_addresses() {
+        // A host resolving to one public + one private IP yields only
+        // the public one to reqwest.
+        let public: SocketAddr = "1.1.1.1:0".parse().unwrap();
+        let private: SocketAddr = "10.0.0.1:0".parse().unwrap();
+        let resolver =
+            SsrfResolver::with_lookup(SsrfPolicy::new(false), fixed_lookup(vec![public, private]));
+        let got: Vec<SocketAddr> = resolver
+            .resolve(name("mixed.example"))
+            .await
+            .unwrap()
+            .collect();
+        assert_eq!(got, vec![public]);
+    }
+
+    #[tokio::test]
+    async fn resolver_all_denied_fails_with_ssrf_marker() {
+        // A host resolving to ONLY private IPs fails with the SSRF
+        // marker (not a generic empty/DNS result).
+        let resolver = SsrfResolver::with_lookup(
+            SsrfPolicy::new(false),
+            fixed_lookup(vec![
+                "10.0.0.1:0".parse().unwrap(),
+                "127.0.0.1:0".parse().unwrap(),
+            ]),
+        );
+        let Err(err) = resolver.resolve(name("internal.example")).await else {
+            panic!("all-denied resolution should error");
+        };
+        assert!(
+            err.to_string().starts_with(SSRF_BLOCK_PREFIX),
+            "expected SSRF marker, got: {err}"
+        );
+    }
+
+    #[tokio::test]
+    async fn resolver_dns_rebinding_second_resolution_denied() {
+        // Simulate rebinding: public IP on the first lookup, private on
+        // the second. The connect-time filter denies the second.
+        let calls = Arc::new(AtomicUsize::new(0));
+        let calls2 = calls.clone();
+        let lookup: LookupFn = Arc::new(move |_host| {
+            let n = calls2.fetch_add(1, Ordering::SeqCst);
+            Box::pin(async move {
+                let addr: SocketAddr = if n == 0 {
+                    "1.1.1.1:0".parse().unwrap()
+                } else {
+                    "127.0.0.1:0".parse().unwrap()
+                };
+                Ok(vec![addr])
+            })
+        });
+        let resolver = SsrfResolver::with_lookup(SsrfPolicy::new(false), lookup);
+
+        // First resolution: public → allowed.
+        let first: Vec<SocketAddr> = resolver
+            .resolve(name("rebind.example"))
+            .await
+            .unwrap()
+            .collect();
+        assert_eq!(first, vec!["1.1.1.1:0".parse::<SocketAddr>().unwrap()]);
+
+        // Second resolution: rebinding returns loopback → denied.
+        let Err(err) = resolver.resolve(name("rebind.example")).await else {
+            panic!("rebound private address must be denied");
+        };
+        assert!(err.to_string().contains("loopback"));
+    }
+
+    #[tokio::test]
+    async fn resolver_empty_resolution_is_not_ssrf() {
+        // Genuine DNS miss (no addresses) returns an empty iterator,
+        // NOT the SSRF marker — reqwest surfaces a normal DNS error.
+        let resolver = SsrfResolver::with_lookup(SsrfPolicy::new(false), fixed_lookup(vec![]));
+        let got: Vec<SocketAddr> = resolver
+            .resolve(name("nxdomain.example"))
+            .await
+            .unwrap()
+            .collect();
+        assert!(got.is_empty());
+    }
+}

crates/manager-core/src/trigger_config.rs

@@ -56,6 +56,11 @@ pub struct TriggerConfig {
     pub dead_letter_retention_days: u32,
     /// abandoned-execution retention before GC, in days. Default 7.
     pub abandoned_retention_days: u32,
+
+    /// Cron scheduler poll cadence, in ms (v1.1.4). Default 30 000 —
+    /// real-world cron precision is per-minute, so a 30s tick is fine.
+    /// Floored at 1s by the scheduler.
+    pub cron_tick_interval_ms: u32,
 }
 
 impl TriggerConfig {
@@ -69,6 +74,7 @@ impl TriggerConfig {
             retry_jitter_pct: 20,
             dead_letter_retention_days: 30,
             abandoned_retention_days: 7,
+            cron_tick_interval_ms: 30_000,
         }
     }
 
@@ -91,6 +97,10 @@ impl TriggerConfig {
             &mut c.abandoned_retention_days,
             "PICLOUD_ABANDONED_EXECUTIONS_RETENTION_DAYS",
         );
+        load_u32(
+            &mut c.cron_tick_interval_ms,
+            "PICLOUD_CRON_TICK_INTERVAL_MS",
+        );
         c
     }
 }
@@ -141,6 +151,7 @@ mod tests {
         assert_eq!(c.retry_jitter_pct, 20);
         assert_eq!(c.dead_letter_retention_days, 30);
         assert_eq!(c.abandoned_retention_days, 7);
+        assert_eq!(c.cron_tick_interval_ms, 30_000);
     }
 
     #[test]

crates/manager-core/src/trigger_repo.rs

@@ -5,7 +5,9 @@
 
 use async_trait::async_trait;
 use chrono::{DateTime, Utc};
-use picloud_shared::{AdminUserId, AppId, KvEventOp, ScriptId, TriggerId};
+use picloud_shared::{
+    AdminUserId, AppId, DocsEventOp, FilesEventOp, KvEventOp, ScriptId, TriggerId,
+};
 use serde::{Deserialize, Serialize};
 use sqlx::PgPool;
 use uuid::Uuid;
@@ -47,7 +49,14 @@ pub struct Trigger {
 #[serde(rename_all = "snake_case")]
 pub enum TriggerKind {
     Kv,
+    Docs,
     DeadLetter,
+    /// v1.1.4.
+    Cron,
+    /// v1.1.5.
+    Files,
+    /// v1.1.5.
+    Pubsub,
 }
 
 impl TriggerKind {
@@ -55,7 +64,11 @@ impl TriggerKind {
     pub const fn as_str(self) -> &'static str {
         match self {
             Self::Kv => "kv",
+            Self::Docs => "docs",
             Self::DeadLetter => "dead_letter",
+            Self::Cron => "cron",
+            Self::Files => "files",
+            Self::Pubsub => "pubsub",
         }
     }
 
@@ -63,7 +76,11 @@ impl TriggerKind {
     pub fn from_wire(s: &str) -> Option<Self> {
         match s {
             "kv" => Some(Self::Kv),
+            "docs" => Some(Self::Docs),
             "dead_letter" => Some(Self::DeadLetter),
+            "cron" => Some(Self::Cron),
+            "files" => Some(Self::Files),
+            "pubsub" => Some(Self::Pubsub),
             _ => None,
         }
     }
@@ -93,6 +110,10 @@ pub enum TriggerDetails {
         collection_glob: String,
         ops: Vec<KvEventOp>,
     },
+    Docs {
+        collection_glob: String,
+        ops: Vec<DocsEventOp>,
+    },
     DeadLetter {
         #[serde(default, skip_serializing_if = "Option::is_none")]
         source_filter: Option<String>,
@@ -101,6 +122,21 @@ pub enum TriggerDetails {
         #[serde(default, skip_serializing_if = "Option::is_none")]
         script_id_filter: Option<ScriptId>,
     },
+    /// v1.1.4. The 6-field cron schedule + IANA timezone the trigger
+    /// fires on, plus the last enqueue time (for dashboard display).
+    Cron {
+        schedule: String,
+        timezone: String,
+        #[serde(default, skip_serializing_if = "Option::is_none")]
+        last_fired_at: Option<DateTime<Utc>>,
+    },
+    /// v1.1.5. Same shape as KV/docs: a collection glob + op subset.
+    Files {
+        collection_glob: String,
+        ops: Vec<FilesEventOp>,
+    },
+    /// v1.1.5. A topic pattern: exact, `<prefix>.*`, or `*`.
+    Pubsub { topic_pattern: String },
 }
 
 /// Create payload for a KV trigger. Defaults applied at the admin
@@ -118,6 +154,20 @@ pub struct CreateKvTrigger {
     pub registered_by_principal: AdminUserId,
 }
 
+/// Create payload for a docs trigger (v1.1.2). Same shape as KV with
+/// `DocsEventOp` ops instead of `KvEventOp`.
+#[derive(Debug, Clone)]
+pub struct CreateDocsTrigger {
+    pub script_id: ScriptId,
+    pub collection_glob: String,
+    pub ops: Vec<DocsEventOp>,
+    pub dispatch_mode: TriggerDispatchMode,
+    pub retry_max_attempts: u32,
+    pub retry_backoff: BackoffShape,
+    pub retry_base_ms: u32,
+    pub registered_by_principal: AdminUserId,
+}
+
 #[derive(Debug, Clone)]
 pub struct CreateDeadLetterTrigger {
     pub script_id: ScriptId,
@@ -127,6 +177,61 @@ pub struct CreateDeadLetterTrigger {
     pub registered_by_principal: AdminUserId,
 }
 
+/// Create payload for a cron trigger (v1.1.4). `schedule` is a 6-field
+/// cron expression and `timezone` an IANA name; both are validated
+/// (by the admin endpoint and defensively by the repo) before insert.
+#[derive(Debug, Clone)]
+pub struct CreateCronTrigger {
+    pub script_id: ScriptId,
+    pub schedule: String,
+    pub timezone: String,
+    pub dispatch_mode: TriggerDispatchMode,
+    pub retry_max_attempts: u32,
+    pub retry_backoff: BackoffShape,
+    pub retry_base_ms: u32,
+    pub registered_by_principal: AdminUserId,
+}
+
+/// Create payload for a files trigger (v1.1.5). Same shape as KV with
+/// `FilesEventOp` ops.
+#[derive(Debug, Clone)]
+pub struct CreateFilesTrigger {
+    pub script_id: ScriptId,
+    pub collection_glob: String,
+    pub ops: Vec<FilesEventOp>,
+    pub dispatch_mode: TriggerDispatchMode,
+    pub retry_max_attempts: u32,
+    pub retry_backoff: BackoffShape,
+    pub retry_base_ms: u32,
+    pub registered_by_principal: AdminUserId,
+}
+
+/// One match for the dispatcher's files trigger fan-out lookup
+/// (v1.1.5). Same shape as `KvTriggerMatch`.
+#[derive(Debug, Clone)]
+pub struct FilesTriggerMatch {
+    pub trigger_id: TriggerId,
+    pub script_id: ScriptId,
+    pub dispatch_mode: TriggerDispatchMode,
+    pub retry_max_attempts: u32,
+    pub retry_backoff: BackoffShape,
+    pub retry_base_ms: u32,
+    pub registered_by_principal: AdminUserId,
+}
+
+/// Create payload for a pubsub trigger (v1.1.5). `topic_pattern` is
+/// validated (exact / `<prefix>.*` / `*`) before insert.
+#[derive(Debug, Clone)]
+pub struct CreatePubsubTrigger {
+    pub script_id: ScriptId,
+    pub topic_pattern: String,
+    pub dispatch_mode: TriggerDispatchMode,
+    pub retry_max_attempts: u32,
+    pub retry_backoff: BackoffShape,
+    pub retry_base_ms: u32,
+    pub registered_by_principal: AdminUserId,
+}
+
 /// One match for the dispatcher's "which KV triggers fire on this
 /// event" lookup. Carries everything the dispatcher needs to construct
 /// the outbox row.
@@ -141,6 +246,19 @@ pub struct KvTriggerMatch {
     pub registered_by_principal: AdminUserId,
 }
 
+/// One match for the dispatcher's docs trigger fan-out lookup (v1.1.2).
+/// Same shape as `KvTriggerMatch`.
+#[derive(Debug, Clone)]
+pub struct DocsTriggerMatch {
+    pub trigger_id: TriggerId,
+    pub script_id: ScriptId,
+    pub dispatch_mode: TriggerDispatchMode,
+    pub retry_max_attempts: u32,
+    pub retry_backoff: BackoffShape,
+    pub retry_base_ms: u32,
+    pub registered_by_principal: AdminUserId,
+}
+
 /// One match for the dispatcher's "which dead-letter triggers fire
 /// on this dead-letter row" lookup.
 #[derive(Debug, Clone)]
@@ -159,12 +277,42 @@ pub trait TriggerRepo: Send + Sync {
         req: CreateKvTrigger,
     ) -> Result<Trigger, TriggerRepoError>;
 
+    /// v1.1.2.
+    async fn create_docs_trigger(
+        &self,
+        app_id: AppId,
+        req: CreateDocsTrigger,
+    ) -> Result<Trigger, TriggerRepoError>;
+
     async fn create_dead_letter_trigger(
         &self,
         app_id: AppId,
         req: CreateDeadLetterTrigger,
     ) -> Result<Trigger, TriggerRepoError>;
 
+    /// v1.1.4. `schedule` + `timezone` are validated before insert; an
+    /// invalid expression or unknown IANA name returns
+    /// `TriggerRepoError::Invalid`.
+    async fn create_cron_trigger(
+        &self,
+        app_id: AppId,
+        req: CreateCronTrigger,
+    ) -> Result<Trigger, TriggerRepoError>;
+
+    /// v1.1.5.
+    async fn create_files_trigger(
+        &self,
+        app_id: AppId,
+        req: CreateFilesTrigger,
+    ) -> Result<Trigger, TriggerRepoError>;
+
+    /// v1.1.5. `topic_pattern` is validated before insert.
+    async fn create_pubsub_trigger(
+        &self,
+        app_id: AppId,
+        req: CreatePubsubTrigger,
+    ) -> Result<Trigger, TriggerRepoError>;
+
     async fn list_for_app(&self, app_id: AppId) -> Result<Vec<Trigger>, TriggerRepoError>;
 
     async fn get(&self, id: TriggerId) -> Result<Option<Trigger>, TriggerRepoError>;
@@ -182,6 +330,26 @@ pub trait TriggerRepo: Send + Sync {
         op: KvEventOp,
     ) -> Result<Vec<KvTriggerMatch>, TriggerRepoError>;
 
+    /// Dispatcher hot path for docs fan-out (v1.1.2). Mirrors the KV
+    /// fan-out logic: pull every enabled docs trigger, filter glob +
+    /// ops in Rust (empty ops array means "any op").
+    async fn list_matching_docs(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        op: DocsEventOp,
+    ) -> Result<Vec<DocsTriggerMatch>, TriggerRepoError>;
+
+    /// Dispatcher hot path for files fan-out (v1.1.5). Mirrors the KV
+    /// fan-out logic: pull every enabled files trigger, filter glob +
+    /// ops in Rust (empty ops array means "any op").
+    async fn list_matching_files(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        op: FilesEventOp,
+    ) -> Result<Vec<FilesTriggerMatch>, TriggerRepoError>;
+
     /// Dispatcher hot path for dead-letter fan-out. Filters: source
     /// (or any-source), originating trigger_id (or any), originating
     /// script_id (or any). Each filter is "match OR is_null".
@@ -276,6 +444,71 @@ impl TriggerRepo for PostgresTriggerRepo {
         })
     }
 
+    async fn create_docs_trigger(
+        &self,
+        app_id: AppId,
+        req: CreateDocsTrigger,
+    ) -> Result<Trigger, TriggerRepoError> {
+        if req.collection_glob.is_empty() {
+            return Err(TriggerRepoError::Invalid(
+                "collection_glob must not be empty".into(),
+            ));
+        }
+        let mut tx = self.pool.begin().await?;
+        let parent: TriggerRow = sqlx::query_as(
+            "INSERT INTO triggers ( \
+                app_id, script_id, kind, enabled, dispatch_mode, \
+                retry_max_attempts, retry_backoff, retry_base_ms, \
+                registered_by_principal \
+             ) VALUES ($1, $2, 'docs', TRUE, $3, $4, $5, $6, $7) \
+             RETURNING id, app_id, script_id, kind, enabled, dispatch_mode, \
+                       retry_max_attempts, retry_backoff, retry_base_ms, \
+                       registered_by_principal, created_at, updated_at",
+        )
+        .bind(app_id.into_inner())
+        .bind(req.script_id.into_inner())
+        .bind(req.dispatch_mode.as_str())
+        .bind(i32::try_from(req.retry_max_attempts).unwrap_or(3))
+        .bind(req.retry_backoff.as_str())
+        .bind(i32::try_from(req.retry_base_ms).unwrap_or(1000))
+        .bind(req.registered_by_principal.into_inner())
+        .fetch_one(&mut *tx)
+        .await?;
+
+        let ops_str: Vec<String> = req.ops.iter().map(|o| o.as_str().to_string()).collect();
+        sqlx::query(
+            "INSERT INTO docs_trigger_details (trigger_id, collection_glob, ops) \
+             VALUES ($1, $2, $3)",
+        )
+        .bind(parent.id)
+        .bind(&req.collection_glob)
+        .bind(&ops_str)
+        .execute(&mut *tx)
+        .await?;
+
+        tx.commit().await?;
+
+        Ok(Trigger {
+            id: parent.id.into(),
+            app_id: parent.app_id.into(),
+            script_id: parent.script_id.into(),
+            kind: TriggerKind::Docs,
+            enabled: parent.enabled,
+            dispatch_mode: dispatch_from_str(&parent.dispatch_mode),
+            retry_max_attempts: u32::try_from(parent.retry_max_attempts).unwrap_or(3),
+            retry_backoff: BackoffShape::from_wire(&parent.retry_backoff)
+                .unwrap_or(BackoffShape::Exponential),
+            retry_base_ms: u32::try_from(parent.retry_base_ms).unwrap_or(1000),
+            registered_by_principal: parent.registered_by_principal.into(),
+            created_at: parent.created_at,
+            updated_at: parent.updated_at,
+            details: TriggerDetails::Docs {
+                collection_glob: req.collection_glob,
+                ops: req.ops,
+            },
+        })
+    }
+
     async fn create_dead_letter_trigger(
         &self,
         app_id: AppId,
@@ -337,6 +570,197 @@ impl TriggerRepo for PostgresTriggerRepo {
         })
     }
 
+    async fn create_cron_trigger(
+        &self,
+        app_id: AppId,
+        req: CreateCronTrigger,
+    ) -> Result<Trigger, TriggerRepoError> {
+        // Defense-in-depth validation (the admin endpoint validates too).
+        crate::cron_scheduler::validate_schedule(&req.schedule)
+            .map_err(|e| TriggerRepoError::Invalid(format!("invalid cron schedule: {e}")))?;
+        crate::cron_scheduler::validate_timezone(&req.timezone)
+            .map_err(|e| TriggerRepoError::Invalid(format!("invalid timezone: {e}")))?;
+
+        let mut tx = self.pool.begin().await?;
+        let parent: TriggerRow = sqlx::query_as(
+            "INSERT INTO triggers ( \
+                app_id, script_id, kind, enabled, dispatch_mode, \
+                retry_max_attempts, retry_backoff, retry_base_ms, \
+                registered_by_principal \
+             ) VALUES ($1, $2, 'cron', TRUE, $3, $4, $5, $6, $7) \
+             RETURNING id, app_id, script_id, kind, enabled, dispatch_mode, \
+                       retry_max_attempts, retry_backoff, retry_base_ms, \
+                       registered_by_principal, created_at, updated_at",
+        )
+        .bind(app_id.into_inner())
+        .bind(req.script_id.into_inner())
+        .bind(req.dispatch_mode.as_str())
+        .bind(i32::try_from(req.retry_max_attempts).unwrap_or(3))
+        .bind(req.retry_backoff.as_str())
+        .bind(i32::try_from(req.retry_base_ms).unwrap_or(1000))
+        .bind(req.registered_by_principal.into_inner())
+        .fetch_one(&mut *tx)
+        .await?;
+
+        sqlx::query(
+            "INSERT INTO cron_trigger_details (trigger_id, schedule, timezone) \
+             VALUES ($1, $2, $3)",
+        )
+        .bind(parent.id)
+        .bind(&req.schedule)
+        .bind(&req.timezone)
+        .execute(&mut *tx)
+        .await?;
+
+        tx.commit().await?;
+
+        Ok(Trigger {
+            id: parent.id.into(),
+            app_id: parent.app_id.into(),
+            script_id: parent.script_id.into(),
+            kind: TriggerKind::Cron,
+            enabled: parent.enabled,
+            dispatch_mode: dispatch_from_str(&parent.dispatch_mode),
+            retry_max_attempts: u32::try_from(parent.retry_max_attempts).unwrap_or(3),
+            retry_backoff: BackoffShape::from_wire(&parent.retry_backoff)
+                .unwrap_or(BackoffShape::Exponential),
+            retry_base_ms: u32::try_from(parent.retry_base_ms).unwrap_or(1000),
+            registered_by_principal: parent.registered_by_principal.into(),
+            created_at: parent.created_at,
+            updated_at: parent.updated_at,
+            details: TriggerDetails::Cron {
+                schedule: req.schedule,
+                timezone: req.timezone,
+                last_fired_at: None,
+            },
+        })
+    }
+
+    async fn create_files_trigger(
+        &self,
+        app_id: AppId,
+        req: CreateFilesTrigger,
+    ) -> Result<Trigger, TriggerRepoError> {
+        if req.collection_glob.is_empty() {
+            return Err(TriggerRepoError::Invalid(
+                "collection_glob must not be empty".into(),
+            ));
+        }
+        let mut tx = self.pool.begin().await?;
+        let parent: TriggerRow = sqlx::query_as(
+            "INSERT INTO triggers ( \
+                app_id, script_id, kind, enabled, dispatch_mode, \
+                retry_max_attempts, retry_backoff, retry_base_ms, \
+                registered_by_principal \
+             ) VALUES ($1, $2, 'files', TRUE, $3, $4, $5, $6, $7) \
+             RETURNING id, app_id, script_id, kind, enabled, dispatch_mode, \
+                       retry_max_attempts, retry_backoff, retry_base_ms, \
+                       registered_by_principal, created_at, updated_at",
+        )
+        .bind(app_id.into_inner())
+        .bind(req.script_id.into_inner())
+        .bind(req.dispatch_mode.as_str())
+        .bind(i32::try_from(req.retry_max_attempts).unwrap_or(3))
+        .bind(req.retry_backoff.as_str())
+        .bind(i32::try_from(req.retry_base_ms).unwrap_or(1000))
+        .bind(req.registered_by_principal.into_inner())
+        .fetch_one(&mut *tx)
+        .await?;
+
+        let ops_str: Vec<String> = req.ops.iter().map(|o| o.as_str().to_string()).collect();
+        sqlx::query(
+            "INSERT INTO files_trigger_details (trigger_id, collection_glob, ops) \
+             VALUES ($1, $2, $3)",
+        )
+        .bind(parent.id)
+        .bind(&req.collection_glob)
+        .bind(&ops_str)
+        .execute(&mut *tx)
+        .await?;
+
+        tx.commit().await?;
+
+        Ok(Trigger {
+            id: parent.id.into(),
+            app_id: parent.app_id.into(),
+            script_id: parent.script_id.into(),
+            kind: TriggerKind::Files,
+            enabled: parent.enabled,
+            dispatch_mode: dispatch_from_str(&parent.dispatch_mode),
+            retry_max_attempts: u32::try_from(parent.retry_max_attempts).unwrap_or(3),
+            retry_backoff: BackoffShape::from_wire(&parent.retry_backoff)
+                .unwrap_or(BackoffShape::Exponential),
+            retry_base_ms: u32::try_from(parent.retry_base_ms).unwrap_or(1000),
+            registered_by_principal: parent.registered_by_principal.into(),
+            created_at: parent.created_at,
+            updated_at: parent.updated_at,
+            details: TriggerDetails::Files {
+                collection_glob: req.collection_glob,
+                ops: req.ops,
+            },
+        })
+    }
+
+    async fn create_pubsub_trigger(
+        &self,
+        app_id: AppId,
+        req: CreatePubsubTrigger,
+    ) -> Result<Trigger, TriggerRepoError> {
+        // Defense-in-depth validation (the admin endpoint validates too).
+        picloud_shared::validate_topic_pattern(&req.topic_pattern)
+            .map_err(TriggerRepoError::Invalid)?;
+
+        let mut tx = self.pool.begin().await?;
+        let parent: TriggerRow = sqlx::query_as(
+            "INSERT INTO triggers ( \
+                app_id, script_id, kind, enabled, dispatch_mode, \
+                retry_max_attempts, retry_backoff, retry_base_ms, \
+                registered_by_principal \
+             ) VALUES ($1, $2, 'pubsub', TRUE, $3, $4, $5, $6, $7) \
+             RETURNING id, app_id, script_id, kind, enabled, dispatch_mode, \
+                       retry_max_attempts, retry_backoff, retry_base_ms, \
+                       registered_by_principal, created_at, updated_at",
+        )
+        .bind(app_id.into_inner())
+        .bind(req.script_id.into_inner())
+        .bind(req.dispatch_mode.as_str())
+        .bind(i32::try_from(req.retry_max_attempts).unwrap_or(3))
+        .bind(req.retry_backoff.as_str())
+        .bind(i32::try_from(req.retry_base_ms).unwrap_or(1000))
+        .bind(req.registered_by_principal.into_inner())
+        .fetch_one(&mut *tx)
+        .await?;
+
+        sqlx::query(
+            "INSERT INTO pubsub_trigger_details (trigger_id, topic_pattern) VALUES ($1, $2)",
+        )
+        .bind(parent.id)
+        .bind(&req.topic_pattern)
+        .execute(&mut *tx)
+        .await?;
+
+        tx.commit().await?;
+
+        Ok(Trigger {
+            id: parent.id.into(),
+            app_id: parent.app_id.into(),
+            script_id: parent.script_id.into(),
+            kind: TriggerKind::Pubsub,
+            enabled: parent.enabled,
+            dispatch_mode: dispatch_from_str(&parent.dispatch_mode),
+            retry_max_attempts: u32::try_from(parent.retry_max_attempts).unwrap_or(3),
+            retry_backoff: BackoffShape::from_wire(&parent.retry_backoff)
+                .unwrap_or(BackoffShape::Exponential),
+            retry_base_ms: u32::try_from(parent.retry_base_ms).unwrap_or(1000),
+            registered_by_principal: parent.registered_by_principal.into(),
+            created_at: parent.created_at,
+            updated_at: parent.updated_at,
+            details: TriggerDetails::Pubsub {
+                topic_pattern: req.topic_pattern,
+            },
+        })
+    }
+
     async fn list_for_app(&self, app_id: AppId) -> Result<Vec<Trigger>, TriggerRepoError> {
         let parents: Vec<TriggerRow> = sqlx::query_as(
             "SELECT id, app_id, script_id, kind, enabled, dispatch_mode, \
@@ -427,6 +851,99 @@ impl TriggerRepo for PostgresTriggerRepo {
         Ok(out)
     }
 
+    async fn list_matching_docs(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        op: DocsEventOp,
+    ) -> Result<Vec<DocsTriggerMatch>, TriggerRepoError> {
+        // Mirrors list_matching_kv: pull every enabled docs trigger,
+        // filter glob + ops in Rust. **Critical**: do NOT push the
+        // ops check into SQL (`WHERE $op = ANY(ops)`) — that would
+        // exclude rows with `ops = '{}'` from the results, breaking
+        // the empty-array-means-any-op semantic.
+        let rows: Vec<KvMatchRow> = sqlx::query_as(
+            "SELECT t.id, t.script_id, t.dispatch_mode, \
+                    t.retry_max_attempts, t.retry_backoff, t.retry_base_ms, \
+                    t.registered_by_principal, \
+                    d.collection_glob, d.ops \
+             FROM triggers t \
+             JOIN docs_trigger_details d ON d.trigger_id = t.id \
+             WHERE t.app_id = $1 AND t.kind = 'docs' AND t.enabled = TRUE",
+        )
+        .bind(app_id.into_inner())
+        .fetch_all(&self.pool)
+        .await?;
+
+        let op_str = op.as_str();
+        let mut out = Vec::new();
+        for r in rows {
+            if !collection_matches(&r.collection_glob, collection) {
+                continue;
+            }
+            let any_op = r.ops.is_empty();
+            if !any_op && !r.ops.iter().any(|o| o == op_str) {
+                continue;
+            }
+            out.push(DocsTriggerMatch {
+                trigger_id: r.id.into(),
+                script_id: r.script_id.into(),
+                dispatch_mode: dispatch_from_str(&r.dispatch_mode),
+                retry_max_attempts: u32::try_from(r.retry_max_attempts).unwrap_or(3),
+                retry_backoff: BackoffShape::from_wire(&r.retry_backoff)
+                    .unwrap_or(BackoffShape::Exponential),
+                retry_base_ms: u32::try_from(r.retry_base_ms).unwrap_or(1000),
+                registered_by_principal: r.registered_by_principal.into(),
+            });
+        }
+        Ok(out)
+    }
+
+    async fn list_matching_files(
+        &self,
+        app_id: AppId,
+        collection: &str,
+        op: FilesEventOp,
+    ) -> Result<Vec<FilesTriggerMatch>, TriggerRepoError> {
+        // Mirrors list_matching_kv: pull every enabled files trigger,
+        // filter glob + ops in Rust (empty ops array means "any op").
+        let rows: Vec<KvMatchRow> = sqlx::query_as(
+            "SELECT t.id, t.script_id, t.dispatch_mode, \
+                    t.retry_max_attempts, t.retry_backoff, t.retry_base_ms, \
+                    t.registered_by_principal, \
+                    d.collection_glob, d.ops \
+             FROM triggers t \
+             JOIN files_trigger_details d ON d.trigger_id = t.id \
+             WHERE t.app_id = $1 AND t.kind = 'files' AND t.enabled = TRUE",
+        )
+        .bind(app_id.into_inner())
+        .fetch_all(&self.pool)
+        .await?;
+
+        let op_str = op.as_str();
+        let mut out = Vec::new();
+        for r in rows {
+            if !collection_matches(&r.collection_glob, collection) {
+                continue;
+            }
+            let any_op = r.ops.is_empty();
+            if !any_op && !r.ops.iter().any(|o| o == op_str) {
+                continue;
+            }
+            out.push(FilesTriggerMatch {
+                trigger_id: r.id.into(),
+                script_id: r.script_id.into(),
+                dispatch_mode: dispatch_from_str(&r.dispatch_mode),
+                retry_max_attempts: u32::try_from(r.retry_max_attempts).unwrap_or(3),
+                retry_backoff: BackoffShape::from_wire(&r.retry_backoff)
+                    .unwrap_or(BackoffShape::Exponential),
+                retry_base_ms: u32::try_from(r.retry_base_ms).unwrap_or(1000),
+                registered_by_principal: r.registered_by_principal.into(),
+            });
+        }
+        Ok(out)
+    }
+
     async fn list_matching_dead_letter(
         &self,
         app_id: AppId,
@@ -463,6 +980,7 @@ impl TriggerRepo for PostgresTriggerRepo {
     }
 }
 
+#[allow(clippy::too_many_lines)]
 async fn hydrate_one(pool: &PgPool, parent: TriggerRow) -> Result<Trigger, TriggerRepoError> {
     let kind = TriggerKind::from_wire(&parent.kind).ok_or_else(|| {
         TriggerRepoError::Invalid(format!("unknown trigger kind {}", parent.kind))
@@ -486,6 +1004,23 @@ async fn hydrate_one(pool: &PgPool, parent: TriggerRow) -> Result<Trigger, Trigg
                 ops,
             }
         }
+        TriggerKind::Docs => {
+            let row: KvDetailRow = sqlx::query_as(
+                "SELECT collection_glob, ops FROM docs_trigger_details WHERE trigger_id = $1",
+            )
+            .bind(parent.id)
+            .fetch_one(pool)
+            .await?;
+            let ops = row
+                .ops
+                .iter()
+                .filter_map(|s| DocsEventOp::from_wire(s))
+                .collect();
+            TriggerDetails::Docs {
+                collection_glob: row.collection_glob,
+                ops,
+            }
+        }
         TriggerKind::DeadLetter => {
             let row: DlDetailRow = sqlx::query_as(
                 "SELECT source_filter, trigger_id_filter, script_id_filter \
@@ -500,6 +1035,48 @@ async fn hydrate_one(pool: &PgPool, parent: TriggerRow) -> Result<Trigger, Trigg
                 script_id_filter: row.script_id_filter.map(Into::into),
             }
         }
+        TriggerKind::Cron => {
+            let row: CronDetailRow = sqlx::query_as(
+                "SELECT schedule, timezone, last_fired_at \
+                 FROM cron_trigger_details WHERE trigger_id = $1",
+            )
+            .bind(parent.id)
+            .fetch_one(pool)
+            .await?;
+            TriggerDetails::Cron {
+                schedule: row.schedule,
+                timezone: row.timezone,
+                last_fired_at: row.last_fired_at,
+            }
+        }
+        TriggerKind::Files => {
+            let row: KvDetailRow = sqlx::query_as(
+                "SELECT collection_glob, ops FROM files_trigger_details WHERE trigger_id = $1",
+            )
+            .bind(parent.id)
+            .fetch_one(pool)
+            .await?;
+            let ops = row
+                .ops
+                .iter()
+                .filter_map(|s| FilesEventOp::from_wire(s))
+                .collect();
+            TriggerDetails::Files {
+                collection_glob: row.collection_glob,
+                ops,
+            }
+        }
+        TriggerKind::Pubsub => {
+            let row: PubsubDetailRow = sqlx::query_as(
+                "SELECT topic_pattern FROM pubsub_trigger_details WHERE trigger_id = $1",
+            )
+            .bind(parent.id)
+            .fetch_one(pool)
+            .await?;
+            TriggerDetails::Pubsub {
+                topic_pattern: row.topic_pattern,
+            }
+        }
     };
 
     Ok(Trigger {
@@ -565,6 +1142,18 @@ struct KvDetailRow {
     ops: Vec<String>,
 }
 
+#[derive(sqlx::FromRow)]
+struct CronDetailRow {
+    schedule: String,
+    timezone: String,
+    last_fired_at: Option<DateTime<Utc>>,
+}
+
+#[derive(sqlx::FromRow)]
+struct PubsubDetailRow {
+    topic_pattern: String,
+}
+
 #[derive(sqlx::FromRow)]
 #[allow(clippy::struct_field_names)]
 struct DlDetailRow {

crates/manager-core/tests/expected_schema.txt

@@ -3,6 +3,16 @@
 
 ## tables
 
+table: abandoned_executions
+  id: uuid NOT NULL default=gen_random_uuid()
+  app_id: uuid NOT NULL
+  outbox_id: uuid NOT NULL
+  script_id: uuid NULL
+  inbox_id: uuid NOT NULL
+  status_code: integer NOT NULL
+  result_summary: text NULL
+  created_at: timestamp with time zone NOT NULL default=now()
+
 table: admin_sessions
   token_hash: text NOT NULL
   user_id: uuid NOT NULL
@@ -61,6 +71,48 @@ table: apps
   created_at: timestamp with time zone NOT NULL default=now()
   updated_at: timestamp with time zone NOT NULL default=now()
 
+table: cron_trigger_details
+  trigger_id: uuid NOT NULL
+  schedule: text NOT NULL
+  timezone: text NOT NULL default='UTC'::text
+  last_fired_at: timestamp with time zone NULL
+
+table: dead_letter_trigger_details
+  trigger_id: uuid NOT NULL
+  source_filter: text NULL
+  trigger_id_filter: uuid NULL
+  script_id_filter: uuid NULL
+
+table: dead_letters
+  id: uuid NOT NULL default=gen_random_uuid()
+  app_id: uuid NOT NULL
+  original_event_id: uuid NOT NULL
+  source: text NOT NULL
+  op: text NOT NULL
+  trigger_id: uuid NULL
+  script_id: uuid NULL
+  payload: jsonb NOT NULL
+  attempt_count: integer NOT NULL
+  first_attempt_at: timestamp with time zone NOT NULL
+  last_attempt_at: timestamp with time zone NOT NULL
+  last_error: text NOT NULL
+  created_at: timestamp with time zone NOT NULL default=now()
+  resolved_at: timestamp with time zone NULL
+  resolution: text NULL
+
+table: docs
+  app_id: uuid NOT NULL
+  collection: text NOT NULL
+  id: uuid NOT NULL
+  data: jsonb NOT NULL
+  created_at: timestamp with time zone NOT NULL default=now()
+  updated_at: timestamp with time zone NOT NULL default=now()
+
+table: docs_trigger_details
+  trigger_id: uuid NOT NULL
+  collection_glob: text NOT NULL
+  ops: ARRAY NOT NULL
+
 table: execution_logs
   id: uuid NOT NULL default=gen_random_uuid()
   script_id: uuid NOT NULL
@@ -76,6 +128,56 @@ table: execution_logs
   created_at: timestamp with time zone NOT NULL default=now()
   app_id: uuid NOT NULL
 
+table: files
+  app_id: uuid NOT NULL
+  collection: text NOT NULL
+  id: uuid NOT NULL
+  name: text NOT NULL
+  content_type: text NOT NULL
+  size_bytes: bigint NOT NULL
+  checksum_sha256: text NOT NULL
+  created_at: timestamp with time zone NOT NULL default=now()
+  updated_at: timestamp with time zone NOT NULL default=now()
+
+table: files_trigger_details
+  trigger_id: uuid NOT NULL
+  collection_glob: text NOT NULL
+  ops: ARRAY NOT NULL
+
+table: kv_entries
+  app_id: uuid NOT NULL
+  collection: text NOT NULL
+  key: text NOT NULL
+  value: jsonb NOT NULL
+  created_at: timestamp with time zone NOT NULL default=now()
+  updated_at: timestamp with time zone NOT NULL default=now()
+
+table: kv_trigger_details
+  trigger_id: uuid NOT NULL
+  collection_glob: text NOT NULL
+  ops: ARRAY NOT NULL
+
+table: outbox
+  id: uuid NOT NULL default=gen_random_uuid()
+  app_id: uuid NOT NULL
+  source_kind: text NOT NULL
+  trigger_id: uuid NULL
+  script_id: uuid NULL
+  reply_to: uuid NULL
+  payload: jsonb NOT NULL
+  origin_principal: uuid NULL
+  trigger_depth: integer NOT NULL default=0
+  root_execution_id: uuid NULL
+  attempt_count: integer NOT NULL default=0
+  next_attempt_at: timestamp with time zone NOT NULL default=now()
+  claimed_at: timestamp with time zone NULL
+  claimed_by: text NULL
+  created_at: timestamp with time zone NOT NULL default=now()
+
+table: pubsub_trigger_details
+  trigger_id: uuid NOT NULL
+  topic_pattern: text NOT NULL
+
 table: routes
   id: uuid NOT NULL default=gen_random_uuid()
   script_id: uuid NOT NULL
@@ -87,6 +189,13 @@ table: routes
   method: text NULL
   created_at: timestamp with time zone NOT NULL default=now()
   app_id: uuid NOT NULL
+  dispatch_mode: text NOT NULL default='sync'::text
+
+table: script_imports
+  app_id: uuid NOT NULL
+  importer_script_id: uuid NOT NULL
+  imported_script_id: uuid NOT NULL
+  created_at: timestamp with time zone NOT NULL default=now()
 
 table: scripts
   id: uuid NOT NULL default=gen_random_uuid()
@@ -100,9 +209,28 @@ table: scripts
   updated_at: timestamp with time zone NOT NULL default=now()
   sandbox: jsonb NOT NULL default='{}'::jsonb
   app_id: uuid NOT NULL
+  kind: text NOT NULL default='endpoint'::text
+
+table: triggers
+  id: uuid NOT NULL default=gen_random_uuid()
+  app_id: uuid NOT NULL
+  script_id: uuid NOT NULL
+  kind: text NOT NULL
+  enabled: boolean NOT NULL default=true
+  dispatch_mode: text NOT NULL default='async'::text
+  retry_max_attempts: integer NOT NULL
+  retry_backoff: text NOT NULL
+  retry_base_ms: integer NOT NULL
+  registered_by_principal: uuid NOT NULL
+  created_at: timestamp with time zone NOT NULL default=now()
+  updated_at: timestamp with time zone NOT NULL default=now()
 
 ## indexes
 
+indexes on abandoned_executions:
+  abandoned_executions_pkey: public.abandoned_executions USING btree (id)
+  idx_abandoned_executions_gc: public.abandoned_executions USING btree (created_at)
+
 indexes on admin_sessions:
   admin_sessions_expiry_idx: public.admin_sessions USING btree (expires_at)
   admin_sessions_pkey: public.admin_sessions USING btree (token_hash)
@@ -135,11 +263,53 @@ indexes on apps:
   apps_pkey: public.apps USING btree (id)
   apps_slug_key: public.apps USING btree (slug)
 
+indexes on cron_trigger_details:
+  cron_trigger_details_pkey: public.cron_trigger_details USING btree (trigger_id)
+  idx_cron_triggers_due: public.cron_trigger_details USING btree (last_fired_at)
+
+indexes on dead_letter_trigger_details:
+  dead_letter_trigger_details_pkey: public.dead_letter_trigger_details USING btree (trigger_id)
+
+indexes on dead_letters:
+  dead_letters_pkey: public.dead_letters USING btree (id)
+  idx_dead_letters_app_unresolved: public.dead_letters USING btree (app_id) WHERE (resolved_at IS NULL)
+  idx_dead_letters_gc: public.dead_letters USING btree (created_at)
+
+indexes on docs:
+  docs_pkey: public.docs USING btree (app_id, collection, id)
+  idx_docs_app_collection: public.docs USING btree (app_id, collection)
+  idx_docs_data_gin: public.docs USING gin (data jsonb_path_ops)
+
+indexes on docs_trigger_details:
+  docs_trigger_details_pkey: public.docs_trigger_details USING btree (trigger_id)
+
 indexes on execution_logs:
   execution_logs_app_id_created_at_idx: public.execution_logs USING btree (app_id, created_at DESC)
   execution_logs_pkey: public.execution_logs USING btree (id)
   execution_logs_script_id_created_at_idx: public.execution_logs USING btree (script_id, created_at DESC)
 
+indexes on files:
+  files_pkey: public.files USING btree (app_id, collection, id)
+  idx_files_app_collection: public.files USING btree (app_id, collection)
+
+indexes on files_trigger_details:
+  files_trigger_details_pkey: public.files_trigger_details USING btree (trigger_id)
+
+indexes on kv_entries:
+  idx_kv_entries_app_collection: public.kv_entries USING btree (app_id, collection)
+  kv_entries_pkey: public.kv_entries USING btree (app_id, collection, key)
+
+indexes on kv_trigger_details:
+  kv_trigger_details_pkey: public.kv_trigger_details USING btree (trigger_id)
+
+indexes on outbox:
+  idx_outbox_app: public.outbox USING btree (app_id)
+  idx_outbox_due: public.outbox USING btree (next_attempt_at) WHERE (claimed_at IS NULL)
+  outbox_pkey: public.outbox USING btree (id)
+
+indexes on pubsub_trigger_details:
+  pubsub_trigger_details_pkey: public.pubsub_trigger_details USING btree (trigger_id)
+
 indexes on routes:
   routes_app_id_idx: public.routes USING btree (app_id)
   routes_lookup_idx: public.routes USING btree (host_kind, host)
@@ -147,13 +317,28 @@ indexes on routes:
   routes_script_id_idx: public.routes USING btree (script_id)
   routes_unique_binding_idx: public.routes USING btree (app_id, host_kind, host, path_kind, path, COALESCE(method, ''::text))
 
+indexes on script_imports:
+  idx_script_imports_app: public.script_imports USING btree (app_id)
+  idx_script_imports_imported: public.script_imports USING btree (imported_script_id)
+  script_imports_pkey: public.script_imports USING btree (importer_script_id, imported_script_id)
+
 indexes on scripts:
+  idx_scripts_app_kind: public.scripts USING btree (app_id, kind)
   scripts_app_id_idx: public.scripts USING btree (app_id)
   scripts_name_uidx: public.scripts USING btree (app_id, lower(name))
   scripts_pkey: public.scripts USING btree (id)
 
+indexes on triggers:
+  idx_triggers_app_kind_enabled: public.triggers USING btree (app_id, kind) WHERE (enabled = true)
+  idx_triggers_app_pubsub_enabled: public.triggers USING btree (app_id, kind) WHERE ((enabled = true) AND (kind = 'pubsub'::text))
+  triggers_pkey: public.triggers USING btree (id)
+
 ## constraints
 
+constraints on abandoned_executions:
+  [FOREIGN KEY] abandoned_executions_app_id_fkey: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
+  [PRIMARY KEY] abandoned_executions_pkey: PRIMARY KEY (id)
+
 constraints on admin_sessions:
   [FOREIGN KEY] admin_sessions_user_id_fkey: FOREIGN KEY (user_id) REFERENCES admin_users(id) ON DELETE CASCADE
   [PRIMARY KEY] admin_sessions_pkey: PRIMARY KEY (token_hash)
@@ -189,25 +374,89 @@ constraints on apps:
   [PRIMARY KEY] apps_pkey: PRIMARY KEY (id)
   [UNIQUE] apps_slug_key: UNIQUE (slug)
 
+constraints on cron_trigger_details:
+  [FOREIGN KEY] cron_trigger_details_trigger_id_fkey: FOREIGN KEY (trigger_id) REFERENCES triggers(id) ON DELETE CASCADE
+  [PRIMARY KEY] cron_trigger_details_pkey: PRIMARY KEY (trigger_id)
+
+constraints on dead_letter_trigger_details:
+  [FOREIGN KEY] dead_letter_trigger_details_trigger_id_fkey: FOREIGN KEY (trigger_id) REFERENCES triggers(id) ON DELETE CASCADE
+  [PRIMARY KEY] dead_letter_trigger_details_pkey: PRIMARY KEY (trigger_id)
+
+constraints on dead_letters:
+  [CHECK] dead_letters_resolution_check: CHECK ((resolution = ANY (ARRAY['replayed'::text, 'ignored'::text, 'handled_by_script'::text, 'handler_failed'::text])))
+  [FOREIGN KEY] dead_letters_app_id_fkey: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
+  [PRIMARY KEY] dead_letters_pkey: PRIMARY KEY (id)
+
+constraints on docs:
+  [FOREIGN KEY] docs_app_id_fkey: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
+  [PRIMARY KEY] docs_pkey: PRIMARY KEY (app_id, collection, id)
+
+constraints on docs_trigger_details:
+  [FOREIGN KEY] docs_trigger_details_trigger_id_fkey: FOREIGN KEY (trigger_id) REFERENCES triggers(id) ON DELETE CASCADE
+  [PRIMARY KEY] docs_trigger_details_pkey: PRIMARY KEY (trigger_id)
+
 constraints on execution_logs:
   [CHECK] execution_logs_status_check: CHECK ((status = ANY (ARRAY['success'::text, 'error'::text, 'timeout'::text, 'budget_exceeded'::text])))
   [FOREIGN KEY] execution_logs_app_id_fk: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
   [FOREIGN KEY] execution_logs_script_id_fkey: FOREIGN KEY (script_id) REFERENCES scripts(id) ON DELETE CASCADE
   [PRIMARY KEY] execution_logs_pkey: PRIMARY KEY (id)
 
+constraints on files:
+  [FOREIGN KEY] files_app_id_fkey: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
+  [PRIMARY KEY] files_pkey: PRIMARY KEY (app_id, collection, id)
+
+constraints on files_trigger_details:
+  [FOREIGN KEY] files_trigger_details_trigger_id_fkey: FOREIGN KEY (trigger_id) REFERENCES triggers(id) ON DELETE CASCADE
+  [PRIMARY KEY] files_trigger_details_pkey: PRIMARY KEY (trigger_id)
+
+constraints on kv_entries:
+  [FOREIGN KEY] kv_entries_app_id_fkey: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
+  [PRIMARY KEY] kv_entries_pkey: PRIMARY KEY (app_id, collection, key)
+
+constraints on kv_trigger_details:
+  [FOREIGN KEY] kv_trigger_details_trigger_id_fkey: FOREIGN KEY (trigger_id) REFERENCES triggers(id) ON DELETE CASCADE
+  [PRIMARY KEY] kv_trigger_details_pkey: PRIMARY KEY (trigger_id)
+
+constraints on outbox:
+  [CHECK] outbox_source_kind_check: CHECK ((source_kind = ANY (ARRAY['http'::text, 'kv'::text, 'dead_letter'::text, 'docs'::text, 'cron'::text, 'files'::text, 'pubsub'::text])))
+  [FOREIGN KEY] outbox_app_id_fkey: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
+  [PRIMARY KEY] outbox_pkey: PRIMARY KEY (id)
+
+constraints on pubsub_trigger_details:
+  [FOREIGN KEY] pubsub_trigger_details_trigger_id_fkey: FOREIGN KEY (trigger_id) REFERENCES triggers(id) ON DELETE CASCADE
+  [PRIMARY KEY] pubsub_trigger_details_pkey: PRIMARY KEY (trigger_id)
+
 constraints on routes:
+  [CHECK] routes_dispatch_mode_check: CHECK ((dispatch_mode = ANY (ARRAY['sync'::text, 'async'::text])))
   [CHECK] routes_host_kind_check: CHECK ((host_kind = ANY (ARRAY['any'::text, 'strict'::text, 'wildcard'::text])))
   [CHECK] routes_path_kind_check: CHECK ((path_kind = ANY (ARRAY['exact'::text, 'prefix'::text, 'param'::text])))
   [FOREIGN KEY] routes_app_id_fk: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
   [FOREIGN KEY] routes_script_id_fkey: FOREIGN KEY (script_id) REFERENCES scripts(id) ON DELETE CASCADE
   [PRIMARY KEY] routes_pkey: PRIMARY KEY (id)
 
+constraints on script_imports:
+  [FOREIGN KEY] script_imports_app_id_fkey: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
+  [FOREIGN KEY] script_imports_imported_script_id_fkey: FOREIGN KEY (imported_script_id) REFERENCES scripts(id) ON DELETE CASCADE
+  [FOREIGN KEY] script_imports_importer_script_id_fkey: FOREIGN KEY (importer_script_id) REFERENCES scripts(id) ON DELETE CASCADE
+  [PRIMARY KEY] script_imports_pkey: PRIMARY KEY (importer_script_id, imported_script_id)
+
 constraints on scripts:
+  [CHECK] scripts_kind_check: CHECK ((kind = ANY (ARRAY['endpoint'::text, 'module'::text])))
   [CHECK] scripts_memory_limit_mb_check: CHECK (((memory_limit_mb > 0) AND (memory_limit_mb <= 2048)))
+  [CHECK] scripts_module_name_shape: CHECK (((kind <> 'module'::text) OR (name ~ '^[a-zA-Z_][a-zA-Z0-9_]{0,63}$'::text)))
   [CHECK] scripts_timeout_seconds_check: CHECK (((timeout_seconds > 0) AND (timeout_seconds <= 300)))
   [FOREIGN KEY] scripts_app_id_fk: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE RESTRICT
   [PRIMARY KEY] scripts_pkey: PRIMARY KEY (id)
 
+constraints on triggers:
+  [CHECK] triggers_dispatch_mode_check: CHECK ((dispatch_mode = ANY (ARRAY['sync'::text, 'async'::text])))
+  [CHECK] triggers_kind_check: CHECK ((kind = ANY (ARRAY['kv'::text, 'dead_letter'::text, 'docs'::text, 'cron'::text, 'files'::text, 'pubsub'::text])))
+  [CHECK] triggers_retry_backoff_check: CHECK ((retry_backoff = ANY (ARRAY['exponential'::text, 'linear'::text, 'constant'::text])))
+  [FOREIGN KEY] triggers_app_id_fkey: FOREIGN KEY (app_id) REFERENCES apps(id) ON DELETE CASCADE
+  [FOREIGN KEY] triggers_registered_by_principal_fkey: FOREIGN KEY (registered_by_principal) REFERENCES admin_users(id) ON DELETE CASCADE
+  [FOREIGN KEY] triggers_script_id_fkey: FOREIGN KEY (script_id) REFERENCES scripts(id) ON DELETE CASCADE
+  [PRIMARY KEY] triggers_pkey: PRIMARY KEY (id)
+
 ## applied migrations
   0001: init
   0002: sandbox
@@ -215,3 +464,17 @@ constraints on scripts:
   0004: admin auth
   0005: apps
   0006: users authz
+  0007: kv
+  0008: triggers
+  0009: outbox
+  0010: dead letters
+  0011: abandoned executions
+  0012: routes dispatch mode
+  0013: docs
+  0014: docs triggers
+  0015: scripts kind
+  0016: script imports
+  0017: cron triggers
+  0018: files
+  0019: files triggers
+  0020: pubsub triggers

crates/manager-core/tests/schema_snapshot.rs

@@ -25,22 +25,46 @@
 //!
 //! Review the resulting diff in the same PR as the new migration.
 //!
-//! Like the orchestrator integration tests, this is `#[ignore]`'d by
-//! default so plain `cargo test --workspace` stays green without
-//! infrastructure.
+//! v1.1.5: this test is no longer `#[ignore]`'d. It runs whenever
+//! `DATABASE_URL` is set (CI wires a `postgres:15` service) and **skips
+//! cleanly** when it's absent, so plain `cargo test --workspace` stays
+//! green on machines without Postgres. Unlike the previous
+//! `#[sqlx::test]` form (which spun up an isolated throwaway database),
+//! it now applies the migrations against the `DATABASE_URL` database
+//! directly — migrations are forward-only and idempotent, and CI's
+//! Postgres is fresh, so the structural dump is identical either way.
 
 use std::fmt::Write as _;
 use std::path::PathBuf;
 
+use sqlx::postgres::PgPoolOptions;
 use sqlx::{PgPool, Row};
 
 const SCHEMA: &str = "public";
 
 const SNAPSHOT_PATH: &str = "tests/expected_schema.txt";
 
-#[ignore = "needs DATABASE_URL pointing at a running Postgres"]
-#[sqlx::test(migrations = "./migrations")]
-async fn schema_after_replay_matches_snapshot(pool: PgPool) {
+#[tokio::test]
+async fn schema_after_replay_matches_snapshot() {
+    // Skip cleanly when DATABASE_URL is unset so `cargo test --workspace`
+    // stays green without Postgres. CI sets it (postgres:15 service).
+    let Ok(url) = std::env::var("DATABASE_URL") else {
+        eprintln!(
+            "schema_snapshot: DATABASE_URL unset — skipping. Set it (e.g. \
+             postgres://picloud:picloud@localhost:5432/picloud) to run this guardrail."
+        );
+        return;
+    };
+    let pool = PgPoolOptions::new()
+        .max_connections(1)
+        .connect(&url)
+        .await
+        .expect("connect to DATABASE_URL");
+    sqlx::migrate!("./migrations")
+        .run(&pool)
+        .await
+        .expect("apply migrations");
+
     let actual = dump_schema(&pool).await;
 
     let snapshot_file = PathBuf::from(env!("CARGO_MANIFEST_DIR")).join(SNAPSHOT_PATH);

crates/orchestrator-core/Cargo.toml

@@ -21,5 +21,10 @@ tracing.workspace = true
 uuid.workspace = true
 chrono.workspace = true
 reqwest.workspace = true
+rhai.workspace = true
 tokio.workspace = true
 urlencoding.workspace = true
+
+# v1.1.3 — top-level script AST cache lives in orchestrator-core's
+# LocalExecutorClient; key is ScriptId, value is `(updated_at, Arc<rhai::AST>)`.
+lru.workspace = true

crates/orchestrator-core/src/api.rs

@@ -129,7 +129,14 @@ where
 
     let timeout = Duration::from_secs(u64::from(script.timeout_seconds));
     let started = Utc::now();
-    let outcome = state.executor.execute(&script.source, req, timeout).await;
+    let identity = crate::client::ScriptIdentity {
+        script_id: script.id,
+        updated_at: script.updated_at,
+    };
+    let outcome = state
+        .executor
+        .execute_with_identity(identity, &script.source, req, timeout)
+        .await;
     let finished = Utc::now();
 
     // Build and dispatch the audit log regardless of outcome. We await

crates/orchestrator-core/src/client.rs

@@ -1,8 +1,12 @@
-use std::sync::Arc;
+use std::num::NonZeroUsize;
+use std::sync::{Arc, Mutex};
 use std::time::Duration;
 
 use async_trait::async_trait;
+use chrono::{DateTime, Utc};
+use lru::LruCache;
 use picloud_executor_core::{Engine, ExecError, ExecRequest, ExecResponse};
+use picloud_shared::ScriptId;
 
 use crate::gate::{AcquireError, ExecutionGate};
 
@@ -11,6 +15,21 @@ use crate::gate::{AcquireError, ExecutionGate};
 /// resource usage independent of misconfigured scripts.
 const HARD_TIMEOUT_CAP: Duration = Duration::from_secs(300);
 
+/// Default capacity for the top-level script AST cache. Override via
+/// `PICLOUD_SCRIPT_CACHE_SIZE`. Sized assuming a few hundred distinct
+/// endpoint scripts per process.
+const DEFAULT_SCRIPT_CACHE_SIZE: usize = 256;
+
+/// Identity used by [`ExecutorClient::execute_with_identity`] to key
+/// the AST cache. `updated_at` is the freshness comparator — an edit
+/// that bumps `scripts.updated_at` invalidates the cached AST on the
+/// next lookup, no explicit pub/sub.
+#[derive(Debug, Clone, Copy)]
+pub struct ScriptIdentity {
+    pub script_id: ScriptId,
+    pub updated_at: DateTime<Utc>,
+}
+
 /// The seam between the orchestrator and the executor.
 ///
 /// Single-node mode plugs in `LocalExecutorClient`, which calls
@@ -25,6 +44,21 @@ pub trait ExecutorClient: Send + Sync {
         req: ExecRequest,
         timeout: Duration,
     ) -> Result<ExecResponse, ExecError>;
+
+    /// v1.1.3: identity-aware variant for caching. Callers that already
+    /// know the script's `(id, updated_at)` should use this so the local
+    /// executor can reuse a compiled `rhai::AST` across invocations.
+    /// Default impl forwards to `execute` so `RemoteExecutorClient` (and
+    /// any future transport) keeps working without bespoke caching.
+    async fn execute_with_identity(
+        &self,
+        _identity: ScriptIdentity,
+        source: &str,
+        req: ExecRequest,
+        timeout: Duration,
+    ) -> Result<ExecResponse, ExecError> {
+        self.execute(source, req, timeout).await
+    }
 }
 
 /// In-process executor — wraps `executor-core::Engine` directly.
@@ -36,15 +70,106 @@ pub trait ExecutorClient: Send + Sync {
 /// Holds an `ExecutionGate` and acquires a permit before `spawn_blocking`
 /// so a script storm can't drain the blocking-thread pool. The permit
 /// drops with the future, returning the slot.
+///
+/// v1.1.3 adds a top-level AST cache keyed by `ScriptId`. On
+/// `execute_with_identity`, the client compares the caller's
+/// `updated_at` against the cached entry's; a match reuses the
+/// `Arc<rhai::AST>` and skips Rhai's parser. A mismatch (or absence)
+/// triggers a fresh `Engine::compile` + replace.
 pub struct LocalExecutorClient {
     engine: Arc<Engine>,
     gate: Arc<ExecutionGate>,
+    /// `(updated_at, Arc<rhai::AST>)` keyed by `ScriptId`. `Mutex`
+    /// because the cache is shared across invocations of this client;
+    /// LRU eviction caps memory growth.
+    script_cache: Arc<Mutex<LruCache<ScriptId, CachedScript>>>,
+}
+
+pub struct CachedScript {
+    pub updated_at: DateTime<Utc>,
+    pub ast: Arc<rhai::AST>,
 }
 
 impl LocalExecutorClient {
     #[must_use]
     pub fn new(engine: Arc<Engine>, gate: Arc<ExecutionGate>) -> Self {
-        Self { engine, gate }
+        let cap = std::env::var("PICLOUD_SCRIPT_CACHE_SIZE")
+            .ok()
+            .and_then(|s| s.parse::<usize>().ok())
+            .unwrap_or(DEFAULT_SCRIPT_CACHE_SIZE);
+        Self::with_script_cache_capacity(engine, gate, cap)
+    }
+
+    /// Explicit capacity for tests that exercise LRU eviction.
+    #[must_use]
+    pub fn with_script_cache_capacity(
+        engine: Arc<Engine>,
+        gate: Arc<ExecutionGate>,
+        cap: usize,
+    ) -> Self {
+        let cap = NonZeroUsize::new(cap.max(1)).expect("max(1) is non-zero");
+        Self {
+            engine,
+            gate,
+            script_cache: Arc::new(Mutex::new(LruCache::new(cap))),
+        }
+    }
+
+    /// Cache lookup with `updated_at` freshness check. Returns the
+    /// cached AST on hit; compiles, inserts, returns the fresh AST on
+    /// miss or stale. Public so tests can introspect the cache.
+    pub fn get_or_compile(
+        &self,
+        identity: ScriptIdentity,
+        source: &str,
+    ) -> Result<Arc<rhai::AST>, ExecError> {
+        {
+            let mut cache = self
+                .script_cache
+                .lock()
+                .expect("script cache lock poisoned");
+            if let Some(cached) = cache.get(&identity.script_id) {
+                if cached.updated_at == identity.updated_at {
+                    tracing::debug!(
+                        target = "picloud::scripts::cache",
+                        script_id = %identity.script_id,
+                        "cache hit"
+                    );
+                    return Ok(cached.ast.clone());
+                }
+                tracing::debug!(
+                    target = "picloud::scripts::cache",
+                    script_id = %identity.script_id,
+                    "cache stale; recompiling"
+                );
+            } else {
+                tracing::debug!(
+                    target = "picloud::scripts::cache",
+                    script_id = %identity.script_id,
+                    "cache miss"
+                );
+            }
+        }
+        let ast = self.engine.compile(source)?;
+        let mut cache = self
+            .script_cache
+            .lock()
+            .expect("script cache lock poisoned");
+        cache.put(
+            identity.script_id,
+            CachedScript {
+                updated_at: identity.updated_at,
+                ast: ast.clone(),
+            },
+        );
+        Ok(ast)
+    }
+
+    /// Shared script-AST cache. Exposed so tests can introspect cache
+    /// state (length / contents) under a Mutex lock.
+    #[must_use]
+    pub fn script_cache(&self) -> &Arc<Mutex<LruCache<ScriptId, CachedScript>>> {
+        &self.script_cache
     }
 }
 
@@ -89,6 +214,39 @@ impl ExecutorClient for LocalExecutorClient {
             Ok(Ok(res)) => res,
         }
     }
+
+    async fn execute_with_identity(
+        &self,
+        identity: ScriptIdentity,
+        source: &str,
+        req: ExecRequest,
+        timeout: Duration,
+    ) -> Result<ExecResponse, ExecError> {
+        let _permit =
+            self.gate
+                .try_acquire()
+                .map_err(
+                    |AcquireError::Overloaded { retry_after_secs }| ExecError::Overloaded {
+                        retry_after_secs,
+                    },
+                )?;
+
+        let ast = self.get_or_compile(identity, source)?;
+
+        let timeout = timeout.min(HARD_TIMEOUT_CAP);
+        let timeout_secs = u32::try_from(timeout.as_secs()).unwrap_or(u32::MAX);
+
+        let engine = self.engine.clone();
+        let join = tokio::task::spawn_blocking(move || engine.execute_ast(&ast, req));
+
+        match tokio::time::timeout(timeout, join).await {
+            Err(_) => Err(ExecError::Timeout(timeout_secs)),
+            Ok(Err(join_err)) => Err(ExecError::Runtime(format!(
+                "execution task panicked: {join_err}"
+            ))),
+            Ok(Ok(res)) => res,
+        }
+    }
 }
 
 /// Remote executor — forwards to a peer executor node over HTTP.
@@ -122,3 +280,131 @@ impl ExecutorClient for RemoteExecutorClient {
         ))
     }
 }
+
+#[cfg(test)]
+mod cache_tests {
+    use super::*;
+    use picloud_executor_core::Limits;
+    use picloud_shared::Services;
+
+    fn engine() -> Arc<Engine> {
+        Arc::new(Engine::new(Limits::default(), Services::default()))
+    }
+
+    fn client_with_cap(cap: usize) -> LocalExecutorClient {
+        LocalExecutorClient::with_script_cache_capacity(
+            engine(),
+            Arc::new(ExecutionGate::new(32)),
+            cap,
+        )
+    }
+
+    fn identity_at(t: DateTime<Utc>) -> ScriptIdentity {
+        ScriptIdentity {
+            script_id: ScriptId::new(),
+            updated_at: t,
+        }
+    }
+
+    #[test]
+    fn cache_hit_when_identity_matches() {
+        let client = client_with_cap(8);
+        let identity = identity_at(Utc::now());
+        let src = "fn f() { 1 }";
+
+        let ast_a = client.get_or_compile(identity, src).unwrap();
+        let ast_b = client.get_or_compile(identity, src).unwrap();
+
+        // Same Arc — cache served the second call without recompiling.
+        assert!(
+            Arc::ptr_eq(&ast_a, &ast_b),
+            "expected identical Arc<AST> from cache hit"
+        );
+    }
+
+    #[test]
+    fn cache_invalidated_when_updated_at_changes() {
+        let client = client_with_cap(8);
+        let script_id = ScriptId::new();
+        let t0 = Utc::now() - chrono::Duration::seconds(10);
+        let t1 = Utc::now();
+
+        let ast_a = client
+            .get_or_compile(
+                ScriptIdentity {
+                    script_id,
+                    updated_at: t0,
+                },
+                "fn f() { 1 }",
+            )
+            .unwrap();
+        let ast_b = client
+            .get_or_compile(
+                ScriptIdentity {
+                    script_id,
+                    updated_at: t1,
+                },
+                "fn f() { 2 }",
+            )
+            .unwrap();
+
+        // Different Arc — cache miss forced recompile.
+        assert!(
+            !Arc::ptr_eq(&ast_a, &ast_b),
+            "expected recompile on updated_at change"
+        );
+    }
+
+    #[test]
+    fn distinct_script_ids_cache_independently() {
+        let client = client_with_cap(8);
+        let now = Utc::now();
+        let a = identity_at(now);
+        let b = identity_at(now);
+        client.get_or_compile(a, "fn x() { 1 }").unwrap();
+        client.get_or_compile(b, "fn x() { 1 }").unwrap();
+
+        let cache = client.script_cache().lock().unwrap();
+        assert_eq!(
+            cache.len(),
+            2,
+            "distinct script_ids should yield two entries"
+        );
+    }
+
+    #[test]
+    fn lru_eviction_caps_cache_size() {
+        // Capacity 1 — every new script evicts the previous.
+        let client = client_with_cap(1);
+        client
+            .get_or_compile(identity_at(Utc::now()), "fn a() { 1 }")
+            .unwrap();
+        client
+            .get_or_compile(identity_at(Utc::now()), "fn b() { 2 }")
+            .unwrap();
+        client
+            .get_or_compile(identity_at(Utc::now()), "fn c() { 3 }")
+            .unwrap();
+        assert_eq!(client.script_cache().lock().unwrap().len(), 1);
+    }
+
+    #[test]
+    fn script_identity_is_copy() {
+        // Copy is load-bearing — many call sites pass it by value.
+        let id = identity_at(Utc::now());
+        let _ = id;
+        let _ = id; // should still be usable
+    }
+
+    #[test]
+    fn compile_error_does_not_poison_cache() {
+        let client = client_with_cap(8);
+        let identity = identity_at(Utc::now());
+        // Bad source — should error and not insert anything.
+        let res = client.get_or_compile(identity, "@@@ not valid rhai @@@");
+        assert!(res.is_err(), "garbage source should fail to compile");
+        // A subsequent good compile under a fresh identity must still work.
+        let good = client.get_or_compile(identity_at(Utc::now()), "fn ok() { 1 }");
+        assert!(good.is_ok());
+    }
+}

crates/orchestrator-core/src/lib.rs

@@ -16,7 +16,7 @@ pub mod resolver;
 pub mod routing;
 
 pub use api::{data_plane_router, user_routes_router, DataPlaneState};
-pub use client::{ExecutorClient, LocalExecutorClient, RemoteExecutorClient};
+pub use client::{ExecutorClient, LocalExecutorClient, RemoteExecutorClient, ScriptIdentity};
 pub use gate::{AcquireError, ExecutionGate};
 pub use inbox::InboxRegistry;
 pub use resolver::{ResolverError, ScriptResolver};

crates/picloud/src/lib.rs

@@ -11,19 +11,20 @@ use axum::{routing::get, Json, Router};
 use picloud_executor_core::{Engine, Limits};
 use picloud_manager_core::{
     admin_router, admins_router, api_keys_router, app_members_router, apps_api, apps_router,
-    attach_principal_if_present, auth_router, compile_routes, dead_letters_router, migrations,
-    require_authenticated, route_admin_router, triggers_router, AbandonedRepo,
-    AdminPrincipalResolver, AdminSessionRepository, AdminState, AdminUserRepository, AdminsState,
-    ApiKeyRepository, ApiKeysState, AppDomainRepository, AppMembersRepository, AppMembersState,
-    AppRepository, AppsState, AuthState, AuthzRepo, DeadLetterRepo, DeadLettersState, Dispatcher,
-    KvServiceImpl, OutboxEventEmitter, OutboxRepo, PostgresAbandonedRepo,
-    PostgresAdminSessionRepository, PostgresAdminUserRepository, PostgresApiKeyRepository,
-    PostgresAppDomainRepository, PostgresAppMembersRepository, PostgresAppRepository,
-    PostgresDeadLetterRepo, PostgresDeadLetterService, PostgresExecutionLogRepository,
-    PostgresExecutionLogSink, PostgresKvRepo, PostgresOutboxRepo, PostgresRouteRepository,
-    PostgresScriptRepository, PostgresTriggerRepo, PrincipalResolver, RepoResolver,
-    RouteAdminState, RouteRepository, SandboxCeiling, ScriptRepository, TriggerConfig, TriggerRepo,
-    TriggersState,
+    attach_principal_if_present, auth_router, compile_routes, dead_letters_router,
+    files_admin_router, migrations, require_authenticated, route_admin_router, triggers_router,
+    AbandonedRepo, AdminPrincipalResolver, AdminSessionRepository, AdminState, AdminUserRepository,
+    AdminsState, ApiKeyRepository, ApiKeysState, AppDomainRepository, AppMembersRepository,
+    AppMembersState, AppRepository, AppsState, AuthState, AuthzRepo, DeadLetterRepo,
+    DeadLettersState, Dispatcher, DocsServiceImpl, FilesAdminState, FilesConfig, FilesServiceImpl,
+    FsFilesRepo, HttpConfig, HttpServiceImpl, KvServiceImpl, OutboxEventEmitter, OutboxRepo,
+    PostgresAbandonedRepo, PostgresAdminSessionRepository, PostgresAdminUserRepository,
+    PostgresApiKeyRepository, PostgresAppDomainRepository, PostgresAppMembersRepository,
+    PostgresAppRepository, PostgresDeadLetterRepo, PostgresDeadLetterService, PostgresDocsRepo,
+    PostgresExecutionLogRepository, PostgresExecutionLogSink, PostgresKvRepo, PostgresOutboxRepo,
+    PostgresPubsubRepo, PostgresRouteRepository, PostgresScriptRepository, PostgresTriggerRepo,
+    PrincipalResolver, PubsubServiceImpl, RepoResolver, RouteAdminState, RouteRepository,
+    SandboxCeiling, ScriptRepository, TriggerConfig, TriggerRepo, TriggersState,
 };
 use picloud_orchestrator_core::routing::{AppDomainTable, RouteTable};
 use picloud_orchestrator_core::{
@@ -31,8 +32,9 @@ use picloud_orchestrator_core::{
     LocalExecutorClient,
 };
 use picloud_shared::{
-    DeadLetterService, ExecutionLogSink, InboxResolver, KvService, OutboxWriter, ScriptValidator,
-    ServiceEventEmitter, Services, API_VERSION, PRODUCT_VERSION, SDK_VERSION, WIRE_VERSION,
+    DeadLetterService, DocsService, ExecutionLogSink, FilesService, HttpService, InboxResolver,
+    KvService, OutboxWriter, PubsubService, ScriptValidator, ServiceEventEmitter, Services,
+    API_VERSION, PRODUCT_VERSION, SDK_VERSION, WIRE_VERSION,
 };
 use sqlx::postgres::PgPoolOptions;
 use sqlx::PgPool;
@@ -119,22 +121,70 @@ pub async fn build_app(pool: PgPool, auth: AuthDeps) -> anyhow::Result<Router> {
     let abandoned_repo: Arc<dyn AbandonedRepo> = Arc::new(PostgresAbandonedRepo::new(pool.clone()));
     let trigger_config = TriggerConfig::from_env();
 
-    // SDK services bundle. v1.1.1 ships the KV store + the
-    // outbox-backed event emitter + the dead-letter service (replay /
-    // resolve).
-    let kv_repo = Arc::new(PostgresKvRepo::new(pool));
+    // SDK services bundle. v1.1.1 added KV + dead-letter; v1.1.2 added
+    // the docs store; v1.1.3 adds the module source backing the Rhai
+    // resolver. All bound services share the outbox-backed event
+    // emitter so KV and docs mutations both fan out through the same
+    // dispatcher.
+    let kv_repo = Arc::new(PostgresKvRepo::new(pool.clone()));
+    let docs_repo = Arc::new(PostgresDocsRepo::new(pool.clone()));
     let events: Arc<dyn ServiceEventEmitter> = Arc::new(OutboxEventEmitter::new(
         trigger_repo.clone(),
         outbox_repo.clone(),
     ));
     let kv: Arc<dyn KvService> =
         Arc::new(KvServiceImpl::new(kv_repo, authz.clone(), events.clone()));
+    let docs: Arc<dyn DocsService> = Arc::new(DocsServiceImpl::new(
+        docs_repo,
+        authz.clone(),
+        events.clone(),
+    ));
     let dl_service: Arc<dyn DeadLetterService> = Arc::new(PostgresDeadLetterService::new(
         dl_repo.clone(),
         outbox_repo.clone(),
         authz.clone(),
     ));
-    let services = Services::new(kv, dl_service.clone(), events);
+    let modules: Arc<dyn picloud_shared::ModuleSource> = Arc::new(
+        picloud_manager_core::PostgresModuleSource::new(pool.clone()),
+    );
+    // v1.1.4 outbound HTTP. The reqwest client is built once here with
+    // the SSRF deny-list resolver. `PICLOUD_HTTP_ALLOW_PRIVATE=true`
+    // disables the deny-list entirely — dev/test only, so warn loudly.
+    let http_config = HttpConfig::from_env();
+    if http_config.allow_private {
+        tracing::warn!(
+            "PICLOUD_HTTP_ALLOW_PRIVATE is set — the outbound-HTTP SSRF deny-list is DISABLED. \
+             Scripts can reach loopback/private/link-local addresses. Do NOT use in production."
+        );
+    }
+    let http: Arc<dyn HttpService> = Arc::new(HttpServiceImpl::new(http_config, authz.clone()));
+    // v1.1.5 filesystem-backed blob storage. Metadata lives in Postgres;
+    // the bytes live on disk under `PICLOUD_FILES_ROOT` (default ./data).
+    let files_config = FilesConfig::from_env();
+    let files_max_size = files_config.max_file_size_bytes;
+    let files_repo = Arc::new(FsFilesRepo::new(pool.clone(), files_config));
+    let files: Arc<dyn FilesService> = Arc::new(FilesServiceImpl::new(
+        files_repo.clone(),
+        authz.clone(),
+        events.clone(),
+        files_max_size,
+    ));
+    // v1.1.5 durable pub/sub. Publishes fan out to matching pubsub
+    // triggers at publish time (one outbox row each), delivered by the
+    // same dispatcher as every other async trigger.
+    let pubsub_repo = Arc::new(PostgresPubsubRepo::new(pool.clone()));
+    let pubsub: Arc<dyn PubsubService> =
+        Arc::new(PubsubServiceImpl::new(pubsub_repo, authz.clone()));
+    let services = Services::new(
+        kv,
+        docs,
+        dl_service.clone(),
+        events,
+        modules,
+        http,
+        files,
+        pubsub,
+    );
     let engine = Arc::new(Engine::new(Limits::default(), services));
 
     // Compile the routes table once at startup; admin writes refresh it.
@@ -206,7 +256,7 @@ pub async fn build_app(pool: PgPool, auth: AuthDeps) -> anyhow::Result<Router> {
     };
     let route_admin = RouteAdminState {
         routes: route_repo.clone(),
-        scripts: Arc::new(PostgresScriptRepoHandle(script_repo)),
+        scripts: Arc::new(PostgresScriptRepoHandle(script_repo.clone())),
         domains: domains_repo.clone(),
         table: route_table.clone(),
         authz: authz.clone(),
@@ -230,10 +280,15 @@ pub async fn build_app(pool: PgPool, auth: AuthDeps) -> anyhow::Result<Router> {
         abandoned_repo.clone(),
         trigger_config.abandoned_retention_days,
     );
+    // v1.1.4: cron scheduler. Polls cron_trigger_details on a tick and
+    // enqueues due triggers into the outbox; the dispatcher above
+    // delivers them like any other async trigger.
+    picloud_manager_core::spawn_cron_scheduler(pool, trigger_config.cron_tick_interval_ms);
     let triggers_state = TriggersState {
         triggers: trigger_repo,
         apps: apps_repo.clone(),
         authz: authz.clone(),
+        scripts: Arc::new(PostgresScriptRepoHandle(script_repo.clone())),
         config: trigger_config,
     };
     let dead_letters_state = DeadLettersState {
@@ -242,6 +297,11 @@ pub async fn build_app(pool: PgPool, auth: AuthDeps) -> anyhow::Result<Router> {
         apps: apps_repo.clone(),
         authz: authz.clone(),
     };
+    let files_admin_state = FilesAdminState {
+        files: files_repo,
+        apps: apps_repo.clone(),
+        authz: authz.clone(),
+    };
     let apps_state = AppsState {
         apps: apps_repo,
         domains: domains_repo,
@@ -284,6 +344,7 @@ pub async fn build_app(pool: PgPool, auth: AuthDeps) -> anyhow::Result<Router> {
         .merge(app_members_router(app_members_state))
         .merge(api_keys_router(api_keys_state))
         .merge(triggers_router(triggers_state))
+        .merge(files_admin_router(files_admin_state))
         .merge(dead_letters_router(dead_letters_state))
         .layer(from_fn_with_state(
             auth_state.clone(),
@@ -410,4 +471,22 @@ impl picloud_manager_core::ScriptRepository for PostgresScriptRepoHandle {
     ) -> Result<(), picloud_manager_core::ScriptRepositoryError> {
         self.0.delete(id).await
     }
+    async fn count_routes_for_script(
+        &self,
+        script_id: picloud_shared::ScriptId,
+    ) -> Result<i64, picloud_manager_core::ScriptRepositoryError> {
+        self.0.count_routes_for_script(script_id).await
+    }
+    async fn count_triggers_for_script(
+        &self,
+        script_id: picloud_shared::ScriptId,
+    ) -> Result<i64, picloud_manager_core::ScriptRepositoryError> {
+        self.0.count_triggers_for_script(script_id).await
+    }
+    async fn list_imports(
+        &self,
+        script_id: picloud_shared::ScriptId,
+    ) -> Result<Vec<picloud_shared::Script>, picloud_manager_core::ScriptRepositoryError> {
+        self.0.list_imports(script_id).await
+    }
 }

crates/picloud/tests/api.rs

@@ -1221,3 +1221,270 @@ async fn execution_errors_are_still_logged(pool: PgPool) {
     assert_eq!(logs[0]["status"], "error");
     assert!(logs[0]["response_body"]["error"].is_string());
 }
+
+// ============================================================================
+// v1.1.3 — Modules: scripts.kind, route + trigger rejection, end-to-end import
+// ============================================================================
+
+#[ignore = "needs DATABASE_URL pointing at a running Postgres"]
+#[sqlx::test(migrations = "../manager-core/migrations")]
+async fn create_script_default_kind_is_endpoint(pool: PgPool) {
+    let (s, app_id) = server_with_app(pool).await;
+    let r = s
+        .post("/api/v1/admin/scripts")
+        .json(&with_app(
+            &app_id,
+            json!({ "name": "default-kind", "source": "1" }),
+        ))
+        .await;
+    r.assert_status(axum::http::StatusCode::CREATED);
+    let body: Value = r.json();
+    assert_eq!(body["kind"], "endpoint");
+}
+
+#[ignore = "needs DATABASE_URL pointing at a running Postgres"]
+#[sqlx::test(migrations = "../manager-core/migrations")]
+async fn create_module_kind_persists(pool: PgPool) {
+    let (s, app_id) = server_with_app(pool).await;
+    let r = s
+        .post("/api/v1/admin/scripts")
+        .json(&with_app(
+            &app_id,
+            json!({
+                "name": "helpers",
+                "kind": "module",
+                "source": "fn add(a, b) { a + b }"
+            }),
+        ))
+        .await;
+    r.assert_status(axum::http::StatusCode::CREATED);
+    let body: Value = r.json();
+    assert_eq!(body["kind"], "module");
+}
+
+#[ignore = "needs DATABASE_URL pointing at a running Postgres"]
+#[sqlx::test(migrations = "../manager-core/migrations")]
+async fn create_module_with_top_level_expr_rejected(pool: PgPool) {
+    let (s, app_id) = server_with_app(pool).await;
+    let r = s
+        .post("/api/v1/admin/scripts")
+        .json(&with_app(
+            &app_id,
+            json!({
+                "name": "badmod",
+                "kind": "module",
+                "source": "42; fn ok() { 1 }"
+            }),
+        ))
+        .await;
+    r.assert_status(axum::http::StatusCode::UNPROCESSABLE_ENTITY);
+    let body: Value = r.json();
+    assert!(body["error"].as_str().unwrap().contains("module"));
+}
+
+#[ignore = "needs DATABASE_URL pointing at a running Postgres"]
+#[sqlx::test(migrations = "../manager-core/migrations")]
+async fn create_module_with_reserved_name_rejected(pool: PgPool) {
+    let (s, app_id) = server_with_app(pool).await;
+    let r = s
+        .post("/api/v1/admin/scripts")
+        .json(&with_app(
+            &app_id,
+            json!({
+                "name": "kv",
+                "kind": "module",
+                "source": "fn ok() { 1 }"
+            }),
+        ))
+        .await;
+    r.assert_status(axum::http::StatusCode::UNPROCESSABLE_ENTITY);
+    let body: Value = r.json();
+    assert!(body["error"].as_str().unwrap().contains("reserved"));
+}
+
+#[ignore = "needs DATABASE_URL pointing at a running Postgres"]
+#[sqlx::test(migrations = "../manager-core/migrations")]
+async fn route_bind_rejects_module(pool: PgPool) {
+    let (s, app_id) = server_with_app(pool).await;
+    let r = s
+        .post("/api/v1/admin/scripts")
+        .json(&with_app(
+            &app_id,
+            json!({
+                "name": "lib",
+                "kind": "module",
+                "source": "fn pong() { 42 }"
+            }),
+        ))
+        .await;
+    r.assert_status(axum::http::StatusCode::CREATED);
+    let body: Value = r.json();
+    let id = body["id"].as_str().unwrap();
+
+    let r = s
+        .post(&format!("/api/v1/admin/scripts/{id}/routes"))
+        .json(&json!({
+            "host_kind": "any",
+            "path_kind": "exact",
+            "path": "/lib"
+        }))
+        .await;
+    r.assert_status(axum::http::StatusCode::UNPROCESSABLE_ENTITY);
+}
+
+#[ignore = "needs DATABASE_URL pointing at a running Postgres"]
+#[sqlx::test(migrations = "../manager-core/migrations")]
+async fn endpoint_imports_module_end_to_end(pool: PgPool) {
+    let (s, app_id) = server_with_app(pool).await;
+
+    // Create a module script.
+    s.post("/api/v1/admin/scripts")
+        .json(&with_app(
+            &app_id,
+            json!({
+                "name": "math",
+                "kind": "module",
+                "source": "fn add(a, b) { a + b }"
+            }),
+        ))
+        .await
+        .assert_status(axum::http::StatusCode::CREATED);
+
+    // Create an endpoint that imports it.
+    let id = create_basic_script(
+        &s,
+        &app_id,
+        "calc",
+        r#"import "math" as m; #{ statusCode: 200, body: m::add(2, 3) }"#,
+    )
+    .await;
+
+    // Bind a route.
+    s.post(&format!("/api/v1/admin/scripts/{id}/routes"))
+        .json(&json!({
+            "host_kind": "any",
+            "path_kind": "exact",
+            "path": "/calc"
+        }))
+        .await
+        .assert_status(axum::http::StatusCode::CREATED);
+
+    // Hit it — the endpoint should consume the module and return 5.
+    let r = s.get("/calc").add_header("host", "localhost").await;
+    r.assert_status_ok();
+    let body: Value = r.json();
+    assert_eq!(body, json!(5));
+}
+
+#[ignore = "needs DATABASE_URL pointing at a running Postgres"]
+#[sqlx::test(migrations = "../manager-core/migrations")]
+async fn module_edit_visible_on_next_invocation(pool: PgPool) {
+    let (s, app_id) = server_with_app(pool).await;
+
+    let lib: Value = s
+        .post("/api/v1/admin/scripts")
+        .json(&with_app(
+            &app_id,
+            json!({
+                "name": "greet",
+                "kind": "module",
+                "source": r"fn say(n) { `hello, ${n}` }"
+            }),
+        ))
+        .await
+        .json();
+    let lib_id = lib["id"].as_str().unwrap();
+
+    let id = create_basic_script(
+        &s,
+        &app_id,
+        "hello",
+        r#"import "greet" as g; #{ statusCode: 200, body: g::say("world") }"#,
+    )
+    .await;
+    s.post(&format!("/api/v1/admin/scripts/{id}/routes"))
+        .json(&json!({
+            "host_kind": "any",
+            "path_kind": "exact",
+            "path": "/hello"
+        }))
+        .await
+        .assert_status(axum::http::StatusCode::CREATED);
+
+    let r1: Value = s.get("/hello").add_header("host", "localhost").await.json();
+    assert_eq!(r1, json!("hello, world"));
+
+    // Edit the module — bump updated_at.
+    s.put(&format!("/api/v1/admin/scripts/{lib_id}"))
+        .json(&json!({ "source": r"fn say(n) { `hi, ${n}` }" }))
+        .await
+        .assert_status_ok();
+
+    // Cache invalidation must surface the new behavior.
+    let r2: Value = s.get("/hello").add_header("host", "localhost").await.json();
+    assert_eq!(r2, json!("hi, world"));
+}
+
+#[ignore = "needs DATABASE_URL pointing at a running Postgres"]
+#[sqlx::test(migrations = "../manager-core/migrations")]
+async fn cross_app_import_blocked(pool: PgPool) {
+    // Two apps each have a module named "helpers" with different
+    // behavior. An endpoint in app A must import A's module, not B's.
+
+    // App A is already created by `server_with_app`. Create app B.
+    let (s, app_a) = server_with_app(pool).await;
+    let app_b: Value = s
+        .post("/api/v1/admin/apps")
+        .json(&json!({ "slug": "appb", "name": "App B" }))
+        .await
+        .json();
+    let app_b_id = app_b["id"].as_str().unwrap();
+
+    // App A's module returns "A". App B's returns "B".
+    s.post("/api/v1/admin/scripts")
+        .json(&with_app(
+            &app_a,
+            json!({
+                "name": "helpers",
+                "kind": "module",
+                "source": r#"fn who() { "A" }"#
+            }),
+        ))
+        .await
+        .assert_status(axum::http::StatusCode::CREATED);
+    s.post("/api/v1/admin/scripts")
+        .json(&with_app(
+            app_b_id,
+            json!({
+                "name": "helpers",
+                "kind": "module",
+                "source": r#"fn who() { "B" }"#
+            }),
+        ))
+        .await
+        .assert_status(axum::http::StatusCode::CREATED);
+
+    // Endpoint in app A imports "helpers" and exposes the result.
+    let id = create_basic_script(
+        &s,
+        &app_a,
+        "who-am-i",
+        r#"import "helpers" as h; #{ statusCode: 200, body: h::who() }"#,
+    )
+    .await;
+    s.post(&format!("/api/v1/admin/scripts/{id}/routes"))
+        .json(&json!({
+            "host_kind": "any",
+            "path_kind": "exact",
+            "path": "/who-am-i"
+        }))
+        .await
+        .assert_status(axum::http::StatusCode::CREATED);
+
+    let r: Value = s
+        .get("/who-am-i")
+        .add_header("host", "localhost")
+        .await
+        .json();
+    assert_eq!(r, json!("A"), "must see app A's module, not app B's");
+}

crates/shared/src/docs.rs

@@ -0,0 +1,259 @@
+//! `DocsService` — the v1.1.2 schemaless document store contract.
+//!
+//! Lives in `picloud-shared` (not `executor-core`) for the same reason
+//! `KvService` does: the Rhai bridge, the manager-core Postgres impl,
+//! and any future in-memory test impl all depend on the same trait
+//! without dragging `executor-core` into `manager-core`'s dep graph.
+//!
+//! Implementations MUST derive every storage `app_id` from `cx.app_id`
+//! — never from a script-passed argument. That is the cross-app
+//! isolation boundary; see `docs/sdk-shape.md`.
+//!
+//! Filter shape (per `docs::find` / `find_one`) is an opaque
+//! `serde_json::Value` at this layer; the manager-core implementation
+//! parses it into a structured DSL with explicit operator allowlist
+//! before touching SQL. Parser errors surface as
+//! `DocsError::InvalidFilter` / `DocsError::UnsupportedOperator` so
+//! scripts get a clear message naming the offending key.
+
+use async_trait::async_trait;
+use chrono::{DateTime, Utc};
+use thiserror::Error;
+use uuid::Uuid;
+
+use crate::SdkCallCx;
+
+/// Server-generated document identifier. Scripts see the `to_string()`
+/// form as a Rhai string; the trait surface keeps the typed `Uuid` so
+/// no implementation accidentally accepts a string-shaped path
+/// parameter from a script.
+pub type DocId = Uuid;
+
+/// One document as returned by `get` / `find` / `find_one`. The
+/// envelope shape (decision D from the v1.1.2 plan): explicit
+/// `id`+`data`+timestamps so user fields and platform metadata can't
+/// alias. Scripts read user fields via `doc.data.<field>`; timestamps
+/// + id are direct children.
+#[derive(Debug, Clone, PartialEq)]
+pub struct DocRow {
+    pub id: DocId,
+    pub data: serde_json::Value,
+    pub created_at: DateTime<Utc>,
+    pub updated_at: DateTime<Utc>,
+}
+
+/// One page of `list`. `next_cursor` is `Some` when more pages exist,
+/// `None` when exhausted. Mirrors `KvListPage`'s shape; the cursor
+/// encoding is implementation-defined (the Postgres impl base64-encodes
+/// the last id).
+#[derive(Debug, Clone)]
+pub struct DocsListPage {
+    pub docs: Vec<DocRow>,
+    pub next_cursor: Option<String>,
+}
+
+/// Collection-scoped CRUD + cursor list + filter-based find.
+///
+/// Method shapes mirror `KvService`'s signature style (each takes
+/// `&SdkCallCx` first non-self). The collection name is passed by
+/// reference; the implementation rejects empty/whitespace-only
+/// collections at the SDK boundary per `docs/sdk-shape.md`.
+///
+/// `find` and `find_one` take the filter as `serde_json::Value` — the
+/// service implementation parses it into a structured AST. Keeping the
+/// trait signature untyped here lets the bridge convert
+/// `Rhai Map → serde_json::Value` and hand it off without dragging the
+/// parser into the shared crate.
+#[async_trait]
+pub trait DocsService: Send + Sync {
+    /// Create a new document with a server-generated UUID. Returns the
+    /// new id so the script can read/update/delete it later. The
+    /// document `data` must be a JSON object.
+    async fn create(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        data: serde_json::Value,
+    ) -> Result<DocId, DocsError>;
+
+    /// Fetch one document by id. Returns `None` for missing — the
+    /// bridge maps that to Rhai's `()`.
+    async fn get(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: DocId,
+    ) -> Result<Option<DocRow>, DocsError>;
+
+    /// Filter-based query. Returns every matching document as a
+    /// `Vec<DocRow>` (empty when no matches). The filter is the
+    /// v1.1.2 query DSL shape — see `manager-core::docs_filter` for
+    /// the parser. Throws `InvalidFilter` / `UnsupportedOperator` on
+    /// parse errors.
+    async fn find(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        filter: serde_json::Value,
+    ) -> Result<Vec<DocRow>, DocsError>;
+
+    /// Single-result variant — equivalent to `find` with `$limit: 1`
+    /// then take-first. Returns `None` when no document matches.
+    async fn find_one(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        filter: serde_json::Value,
+    ) -> Result<Option<DocRow>, DocsError>;
+
+    /// Full document replace. v1.1.2 has no partial-update DSL —
+    /// scripts that want partial update do `get + modify + update`.
+    /// Returns `DocsError::NotFound` if no such doc; otherwise emits
+    /// an `update` ServiceEvent with `prev_data` and `data`.
+    async fn update(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: DocId,
+        data: serde_json::Value,
+    ) -> Result<(), DocsError>;
+
+    /// Delete by id. Returns `bool was-present` (matches the `delete`
+    /// shape of every v1.1.x service). Emits a `delete` ServiceEvent
+    /// with `prev_data: Some(deleted_doc.data)` when the doc existed.
+    async fn delete(&self, cx: &SdkCallCx, collection: &str, id: DocId) -> Result<bool, DocsError>;
+
+    /// Cursor-paginated listing of every doc in the collection,
+    /// ordered by `id ASC` for stable cursor encoding. `None` cursor
+    /// starts from the beginning. Implementations cap `limit` at a
+    /// reasonable ceiling internally.
+    async fn list(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        cursor: Option<&str>,
+        limit: u32,
+    ) -> Result<DocsListPage, DocsError>;
+}
+
+/// Stub for tests that build a `Services` bundle without spinning up
+/// Postgres. Every call returns `DocsError::Backend("...")` so
+/// accidental docs use surfaces clearly. Mirrors `NoopKvService`.
+#[derive(Debug, Default, Clone, Copy)]
+pub struct NoopDocsService;
+
+#[async_trait]
+impl DocsService for NoopDocsService {
+    async fn create(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _data: serde_json::Value,
+    ) -> Result<DocId, DocsError> {
+        Err(DocsError::Backend("docs is not wired in".into()))
+    }
+
+    async fn get(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _id: DocId,
+    ) -> Result<Option<DocRow>, DocsError> {
+        Err(DocsError::Backend("docs is not wired in".into()))
+    }
+
+    async fn find(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _filter: serde_json::Value,
+    ) -> Result<Vec<DocRow>, DocsError> {
+        Err(DocsError::Backend("docs is not wired in".into()))
+    }
+
+    async fn find_one(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _filter: serde_json::Value,
+    ) -> Result<Option<DocRow>, DocsError> {
+        Err(DocsError::Backend("docs is not wired in".into()))
+    }
+
+    async fn update(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _id: DocId,
+        _data: serde_json::Value,
+    ) -> Result<(), DocsError> {
+        Err(DocsError::Backend("docs is not wired in".into()))
+    }
+
+    async fn delete(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _id: DocId,
+    ) -> Result<bool, DocsError> {
+        Err(DocsError::Backend("docs is not wired in".into()))
+    }
+
+    async fn list(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _cursor: Option<&str>,
+        _limit: u32,
+    ) -> Result<DocsListPage, DocsError> {
+        Err(DocsError::Backend("docs is not wired in".into()))
+    }
+}
+
+/// Failure modes surfaced to the Rhai bridge. The bridge converts each
+/// to a Rhai runtime error string; the discriminants exist so internal
+/// callers (admin endpoints, tests) can react more precisely.
+#[derive(Debug, Error)]
+pub enum DocsError {
+    /// Empty collection name; rejected at the SDK boundary per
+    /// `docs/sdk-shape.md`.
+    #[error("collection name must not be empty")]
+    InvalidCollection,
+
+    /// `create`/`update` was handed a non-object JSON value (data must
+    /// be a JSON object so it can be navigated by field paths in
+    /// queries).
+    #[error("document data must be a JSON object")]
+    InvalidData,
+
+    /// Parser rejected the filter — bad path syntax, malformed
+    /// operator value, multi-field `$sort`, etc. The string is the
+    /// script-visible message; it becomes part of the SDK contract
+    /// once a script depends on it.
+    #[error("invalid filter: {0}")]
+    InvalidFilter(String),
+
+    /// Filter used an operator that's not in the v1.1.2 allowlist
+    /// (`$or`, `$regex`, `$exists`, …). String includes the offending
+    /// operator name + v1.2 pointer.
+    #[error("unsupported operator: {0}")]
+    UnsupportedOperator(String),
+
+    /// `update` / `delete` target id does not exist. (`delete` returns
+    /// `Ok(false)` for "missing"; this variant is for `update` and any
+    /// future delete-must-exist callers.)
+    #[error("document not found")]
+    NotFound,
+
+    /// Caller principal lacked the required capability. Only raised
+    /// when `cx.principal.is_some()` — scripts running with
+    /// `principal: None` (public HTTP) operate under script-as-gate
+    /// semantics and skip the capability check.
+    #[error("forbidden")]
+    Forbidden,
+
+    /// Anything else — Postgres unavailable, serialization failure,
+    /// etc. The string is safe to surface to a script.
+    #[error("docs backend error: {0}")]
+    Backend(String),
+}

crates/shared/src/files.rs

@@ -0,0 +1,339 @@
+//! `FilesService` — the v1.1.5 filesystem-backed blob store contract.
+//!
+//! Lives in `picloud-shared` (not `executor-core`) so the Rhai bridge,
+//! the manager-core filesystem+Postgres impl, and any in-memory test
+//! impl can all depend on the same trait without dragging
+//! `executor-core` into a Postgres or filesystem dependency.
+//!
+//! Implementations MUST derive every storage `app_id` from `cx.app_id`
+//! — never from a script-passed argument. That is the cross-app
+//! isolation boundary; see `docs/sdk-shape.md`.
+//!
+//! `FilesService` is collection-scoped: scripts get a handle via
+//! `files::collection(name)` and call
+//! `create`/`head`/`get`/`update`/`delete`/`list` on it. The blob bytes
+//! never travel through Postgres or through trigger payloads — the row
+//! is metadata + a SHA-256 checksum; the bytes live on the filesystem.
+
+use async_trait::async_trait;
+use chrono::{DateTime, Utc};
+use serde::{Deserialize, Serialize};
+use thiserror::Error;
+use uuid::Uuid;
+
+use crate::SdkCallCx;
+
+/// POSIX-portable filename cap (255 bytes).
+pub const MAX_FILE_NAME_BYTES: usize = 255;
+/// RFC 6838 puts a reasonable media-type ceiling around 127 chars.
+pub const MAX_CONTENT_TYPE_BYTES: usize = 127;
+
+/// Payload for `create` — a brand-new blob. The id is server-generated
+/// (a UUID); scripts never supply it.
+#[derive(Debug, Clone)]
+pub struct NewFile {
+    pub name: String,
+    pub content_type: String,
+    pub data: Vec<u8>,
+}
+
+/// Payload for `update` — replacement bytes plus optional metadata. If
+/// `name` / `content_type` are `None` the prior values are kept.
+#[derive(Debug, Clone)]
+pub struct FileUpdate {
+    pub data: Vec<u8>,
+    pub name: Option<String>,
+    pub content_type: Option<String>,
+}
+
+/// File metadata as scripts and triggers see it. Serialized into
+/// `ServiceEvent.payload` (the blob bytes are NOT included — files are
+/// too big to ship through trigger payloads), and surfaced to Rhai by
+/// `head` / `list`.
+#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
+pub struct FileMeta {
+    pub id: Uuid,
+    pub collection: String,
+    pub name: String,
+    pub content_type: String,
+    pub size: u64,
+    /// Lowercase hex SHA-256 of the content.
+    pub checksum: String,
+    pub created_at: DateTime<Utc>,
+    pub updated_at: DateTime<Utc>,
+}
+
+/// One page of file metadata from `FilesService::list`. `next_cursor`
+/// is `Some` when more pages exist, `None` when exhausted.
+#[derive(Debug, Clone)]
+pub struct FilesListPage {
+    pub files: Vec<FileMeta>,
+    pub next_cursor: Option<String>,
+}
+
+#[async_trait]
+pub trait FilesService: Send + Sync {
+    /// Create a new blob; returns its server-generated id. Throws on a
+    /// missing required field, an over-limit blob, or an invalid
+    /// collection name.
+    async fn create(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        new: NewFile,
+    ) -> Result<Uuid, FilesError>;
+
+    /// Metadata only — no body read. `None` if the file is missing.
+    async fn head(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: &str,
+    ) -> Result<Option<FileMeta>, FilesError>;
+
+    /// Full content. `None` if missing. Verifies the stored checksum
+    /// against the bytes on disk and returns `FilesError::Corrupted`
+    /// when they diverge.
+    async fn get(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: &str,
+    ) -> Result<Option<Vec<u8>>, FilesError>;
+
+    /// Replace content (and optionally metadata). Throws `NotFound`
+    /// when the file doesn't exist.
+    async fn update(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        id: &str,
+        upd: FileUpdate,
+    ) -> Result<(), FilesError>;
+
+    /// Delete by id; returns whether the file was present.
+    async fn delete(&self, cx: &SdkCallCx, collection: &str, id: &str) -> Result<bool, FilesError>;
+
+    /// Cursor-paginated metadata listing (same shape as KV's list).
+    async fn list(
+        &self,
+        cx: &SdkCallCx,
+        collection: &str,
+        cursor: Option<&str>,
+        limit: u32,
+    ) -> Result<FilesListPage, FilesError>;
+}
+
+/// Failure modes surfaced to the Rhai bridge. The bridge converts each
+/// to a Rhai runtime error string; the discriminants exist so internal
+/// callers (admin endpoints, tests) can react more precisely.
+#[derive(Debug, Error)]
+pub enum FilesError {
+    /// Empty collection name, or one containing a path separator / `..`
+    /// / NUL — rejected at the SDK boundary per `docs/sdk-shape.md`.
+    #[error("invalid collection name: {0}")]
+    InvalidCollection(String),
+
+    /// A required field on `create` was missing or empty. The string
+    /// names the field (`name` / `content_type` / `data`).
+    #[error("missing required field: {0}")]
+    MissingField(&'static str),
+
+    /// Blob exceeds the per-file size cap (default 100 MB,
+    /// `PICLOUD_FILES_MAX_FILE_SIZE_BYTES`).
+    #[error("file too large: {size} bytes exceeds limit of {limit} bytes")]
+    TooLarge { size: usize, limit: usize },
+
+    /// Filename exceeds `MAX_FILE_NAME_BYTES`.
+    #[error("file name too long: {0} bytes exceeds 255")]
+    NameTooLong(usize),
+
+    /// Content-type exceeds `MAX_CONTENT_TYPE_BYTES`.
+    #[error("content_type too long: {0} bytes exceeds 127")]
+    ContentTypeTooLong(usize),
+
+    /// `update` on a non-existent file.
+    #[error("file not found")]
+    NotFound,
+
+    /// The bytes on disk no longer match the stored checksum — the
+    /// filesystem corrupted or a backup was misconfigured. The operator
+    /// decides what to do with the metadata-vs-bytes mismatch; the repo
+    /// does NOT auto-delete.
+    #[error("file content corrupted (checksum mismatch)")]
+    Corrupted,
+
+    /// Caller principal lacked the required capability. Only raised when
+    /// `cx.principal.is_some()` — scripts running with `principal: None`
+    /// (public HTTP) operate under script-as-gate semantics and skip
+    /// the capability check.
+    #[error("forbidden")]
+    Forbidden,
+
+    /// Anything else — Postgres unavailable, filesystem I/O error, etc.
+    #[error("files backend error: {0}")]
+    Backend(String),
+}
+
+impl NewFile {
+    /// Validate required fields + length caps at the SDK boundary.
+    /// `data` must be non-empty (v1.1.5 treats an empty blob as a
+    /// missing `data` field — see HANDBACK §7).
+    ///
+    /// # Errors
+    ///
+    /// Returns the field-specific [`FilesError`] for the first failing
+    /// check.
+    pub fn validate(&self, max_size: usize) -> Result<(), FilesError> {
+        if self.name.trim().is_empty() {
+            return Err(FilesError::MissingField("name"));
+        }
+        if self.content_type.trim().is_empty() {
+            return Err(FilesError::MissingField("content_type"));
+        }
+        if self.data.is_empty() {
+            return Err(FilesError::MissingField("data"));
+        }
+        if self.name.len() > MAX_FILE_NAME_BYTES {
+            return Err(FilesError::NameTooLong(self.name.len()));
+        }
+        if self.content_type.len() > MAX_CONTENT_TYPE_BYTES {
+            return Err(FilesError::ContentTypeTooLong(self.content_type.len()));
+        }
+        if self.data.len() > max_size {
+            return Err(FilesError::TooLarge {
+                size: self.data.len(),
+                limit: max_size,
+            });
+        }
+        Ok(())
+    }
+}
+
+impl FileUpdate {
+    /// Validate the replacement bytes + any supplied metadata.
+    ///
+    /// # Errors
+    ///
+    /// Returns the field-specific [`FilesError`] for the first failing
+    /// check.
+    pub fn validate(&self, max_size: usize) -> Result<(), FilesError> {
+        if self.data.is_empty() {
+            return Err(FilesError::MissingField("data"));
+        }
+        if let Some(name) = &self.name {
+            if name.trim().is_empty() {
+                return Err(FilesError::MissingField("name"));
+            }
+            if name.len() > MAX_FILE_NAME_BYTES {
+                return Err(FilesError::NameTooLong(name.len()));
+            }
+        }
+        if let Some(ct) = &self.content_type {
+            if ct.trim().is_empty() {
+                return Err(FilesError::MissingField("content_type"));
+            }
+            if ct.len() > MAX_CONTENT_TYPE_BYTES {
+                return Err(FilesError::ContentTypeTooLong(ct.len()));
+            }
+        }
+        if self.data.len() > max_size {
+            return Err(FilesError::TooLarge {
+                size: self.data.len(),
+                limit: max_size,
+            });
+        }
+        Ok(())
+    }
+}
+
+/// Reject a collection name that is empty or could escape the per-app
+/// files tree. UUID-shaped ids never produce traversal paths, but
+/// collection names come from scripts so they're validated defensively
+/// at both the SDK boundary and the repo.
+///
+/// # Errors
+///
+/// Returns [`FilesError::InvalidCollection`] when the name is empty or
+/// contains `/`, `\`, `..`, or a NUL byte.
+pub fn validate_collection(collection: &str) -> Result<(), FilesError> {
+    if collection.is_empty() {
+        return Err(FilesError::InvalidCollection("must not be empty".into()));
+    }
+    if collection.contains('/')
+        || collection.contains('\\')
+        || collection.contains("..")
+        || collection.contains('\0')
+    {
+        return Err(FilesError::InvalidCollection(format!(
+            "collection {collection:?} must not contain '/', '\\', '..', or NUL"
+        )));
+    }
+    Ok(())
+}
+
+/// Stub used by the test harness so executor-core integration tests
+/// (which don't touch files) can construct a `Services` bundle without
+/// a filesystem or Postgres. Every call returns
+/// `FilesError::Backend("...")` so accidental use surfaces clearly.
+#[derive(Debug, Default, Clone, Copy)]
+pub struct NoopFilesService;
+
+#[async_trait]
+impl FilesService for NoopFilesService {
+    async fn create(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _new: NewFile,
+    ) -> Result<Uuid, FilesError> {
+        Err(FilesError::Backend("files is not wired in".into()))
+    }
+
+    async fn head(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _id: &str,
+    ) -> Result<Option<FileMeta>, FilesError> {
+        Err(FilesError::Backend("files is not wired in".into()))
+    }
+
+    async fn get(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _id: &str,
+    ) -> Result<Option<Vec<u8>>, FilesError> {
+        Err(FilesError::Backend("files is not wired in".into()))
+    }
+
+    async fn update(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _id: &str,
+        _upd: FileUpdate,
+    ) -> Result<(), FilesError> {
+        Err(FilesError::Backend("files is not wired in".into()))
+    }
+
+    async fn delete(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _id: &str,
+    ) -> Result<bool, FilesError> {
+        Err(FilesError::Backend("files is not wired in".into()))
+    }
+
+    async fn list(
+        &self,
+        _cx: &SdkCallCx,
+        _collection: &str,
+        _cursor: Option<&str>,
+        _limit: u32,
+    ) -> Result<FilesListPage, FilesError> {
+        Err(FilesError::Backend("files is not wired in".into()))
+    }
+}

crates/shared/src/http.rs

@@ -0,0 +1,137 @@
+//! `HttpService` — the v1.1.4 outbound-HTTP contract.
+//!
+//! Lives in `picloud-shared` (not `executor-core` or `manager-core`)
+//! so the Rhai bridge and the manager-core reqwest-backed impl can both
+//! depend on the same trait without dragging `executor-core` into
+//! `manager-core`'s dep graph — mirrors [`crate::kv`].
+//!
+//! Unlike KV/docs, `http::*` has no app-scoped data, so there is no
+//! cross-app isolation boundary to enforce here. `cx.app_id` is still
+//! forwarded for audit-log attribution and (future, v1.2) per-app rate
+//! limits. The load-bearing security mechanism is the SSRF deny-list
+//! applied to the *resolved IP* — that lives in the manager-core impl,
+//! not in this contract.
+//!
+//! Body encoding + per-method dispatch happen in the Rhai bridge before
+//! the request reaches this trait: the service receives an already-
+//! encoded body plus a `content_type`, so the impl stays a thin
+//! transport layer.
+
+use std::collections::BTreeMap;
+
+use async_trait::async_trait;
+use thiserror::Error;
+
+use crate::SdkCallCx;
+
+/// A fully-resolved outbound request. The bridge builds this from the
+/// script-facing `(url, body, opts)` arguments; the service backend
+/// turns it into a real network call.
+#[derive(Debug, Clone)]
+pub struct HttpRequest {
+    /// Uppercased HTTP method (`GET`, `POST`, …). The escape-hatch
+    /// `http::request(method, …)` lets scripts pass arbitrary methods,
+    /// so the impl validates this rather than the bridge.
+    pub method: String,
+    pub url: String,
+    /// Caller-supplied headers, merged into the request. Header names
+    /// are case-insensitive on the wire; stored verbatim here.
+    pub headers: BTreeMap<String, String>,
+    /// Already-encoded body. `None` means no body (GET/HEAD, or an
+    /// explicit `()` body).
+    pub body: Option<Vec<u8>>,
+    /// Content-Type the bridge chose for `body` (e.g.
+    /// `application/json`). Ignored when the caller set their own
+    /// `Content-Type` header. `None` when there is no body.
+    pub content_type: Option<String>,
+    /// Total request budget in ms (already clamped to the 60s ceiling
+    /// by the bridge).
+    pub timeout_ms: u32,
+    pub follow_redirects: bool,
+    /// Max redirects to follow (already clamped to 10 by the bridge).
+    pub max_redirects: u32,
+    /// Script id for the default `User-Agent` and audit attribution.
+    /// `None` when unavailable (the bridge always sets it from
+    /// `cx`-adjacent context, but the field stays optional so the
+    /// trait isn't coupled to how the id is sourced).
+    pub script_id: Option<String>,
+}
+
+/// The response shape the bridge turns into a Rhai map. JSON parsing of
+/// `body_raw` happens in the bridge (it needs the Rhai value types), so
+/// the service returns only the raw string + lowercased headers.
+#[derive(Debug, Clone)]
+pub struct HttpResponse {
+    pub status: u16,
+    /// Header names lowercased (per the documented response shape).
+    pub headers: BTreeMap<String, String>,
+    pub body_raw: String,
+}
+
+/// Failure modes surfaced to the Rhai bridge. The bridge prefixes each
+/// `Display` string with `"http: "`. **None of these may leak the
+/// resolved IP** — the SSRF reason is a CIDR-category label only.
+#[derive(Debug, Error)]
+pub enum HttpError {
+    /// Caller principal lacked `AppHttpRequest`. Only raised when
+    /// `cx.principal.is_some()`; public-HTTP scripts skip the check.
+    #[error("forbidden")]
+    Forbidden,
+
+    /// URL failed to parse, or carried no host.
+    #[error("invalid url: {0}")]
+    InvalidUrl(String),
+
+    /// Scheme other than http/https (file, ftp, gopher, …).
+    #[error("scheme not allowed: {0}")]
+    BlockedScheme(String),
+
+    /// Destination port is on the explicit block list (22, 25, 465, 587).
+    #[error("port not allowed: {0}")]
+    BlockedPort(u16),
+
+    /// Resolved IP hit the SSRF deny-list. `reason` is a CIDR-category
+    /// label (e.g. "loopback", "private", "link-local") — never the IP.
+    #[error("blocked by SSRF policy: {0}")]
+    Ssrf(String),
+
+    /// The request exceeded the wall-clock budget.
+    #[error("request timed out")]
+    Timeout,
+
+    /// Request or response body exceeded the configured size cap.
+    /// `which` is `"request"` or `"response"`.
+    #[error("{0} body exceeds size limit")]
+    BodyTooLarge(&'static str),
+
+    /// DNS / connect / TLS failure. The message is generic and MUST NOT
+    /// contain the resolved IP.
+    #[error("{0}")]
+    Network(String),
+
+    /// Anything else the impl wants to surface (still safe to show a
+    /// script).
+    #[error("{0}")]
+    Backend(String),
+}
+
+/// Stub used by the executor-core test harness so engine integration
+/// tests (which don't make real network calls) can construct a
+/// `Services` bundle. Every call errors so accidental use surfaces.
+#[derive(Debug, Default, Clone, Copy)]
+pub struct NoopHttpService;
+
+#[async_trait]
+impl HttpService for NoopHttpService {
+    async fn request(&self, _cx: &SdkCallCx, _req: HttpRequest) -> Result<HttpResponse, HttpError> {
+        Err(HttpError::Network("http is not wired in".into()))
+    }
+}
+
+/// Outbound-HTTP contract. A single generic `request` method funnels
+/// every verb (`get`/`post`/…/`request`); the bridge maps the
+/// script-facing surface onto it.
+#[async_trait]
+pub trait HttpService: Send + Sync {
+    async fn request(&self, cx: &SdkCallCx, req: HttpRequest) -> Result<HttpResponse, HttpError>;
+}

crates/shared/src/lib.rs

@@ -7,15 +7,20 @@
 pub mod app;
 pub mod auth;
 pub mod dead_letters;
+pub mod docs;
 pub mod error;
 pub mod events;
 pub mod exec_summary;
 pub mod execution_log;
+pub mod files;
+pub mod http;
 pub mod ids;
 pub mod inbox;
 pub mod kv;
 pub mod log_sink;
+pub mod modules;
 pub mod outbox_writer;
+pub mod pubsub;
 pub mod route;
 pub mod sandbox;
 pub mod script;
@@ -28,22 +33,34 @@ pub mod version;
 pub use app::{App, AppDomain, DomainShape};
 pub use auth::{AppRole, InstanceRole, Principal, Scope, UserId};
 pub use dead_letters::{DeadLetterError, DeadLetterId, DeadLetterService, NoopDeadLetterService};
+pub use docs::{DocId, DocRow, DocsError, DocsListPage, DocsService, NoopDocsService};
 pub use error::Error;
 pub use events::{EmitError, NoopEventEmitter, ServiceEvent, ServiceEventEmitter};
 pub use exec_summary::ExecResponseSummary;
 pub use execution_log::{ExecutionLog, ExecutionStatus};
+pub use files::{
+    validate_collection as validate_files_collection, FileMeta, FileUpdate, FilesError,
+    FilesListPage, FilesService, NewFile, NoopFilesService,
+};
+pub use http::{HttpError, HttpRequest, HttpResponse, HttpService, NoopHttpService};
 pub use ids::{AdminUserId, ApiKeyId, AppId, ExecutionId, RequestId, ScriptId, TriggerId};
 pub use inbox::{
     InboxDeliveryOutcome, InboxFailureKind, InboxResolver, InboxResult, NoopInboxResolver,
 };
 pub use kv::{KvError, KvListPage, KvService, NoopKvService};
 pub use log_sink::{ExecutionLogSink, LogSinkError};
+pub use modules::{ModuleScript, ModuleSource, ModuleSourceError, NoopModuleSource};
 pub use outbox_writer::{HttpDispatchPayload, NewHttpOutbox, OutboxWriter, OutboxWriterError};
+pub use pubsub::{
+    topic_matches, validate_topic_pattern, NoopPubsubService, PubsubError, PubsubService,
+};
 pub use route::{DispatchMode, HostKind, PathKind, Route};
 pub use sandbox::ScriptSandbox;
-pub use script::Script;
+pub use script::{Script, ScriptKind};
 pub use sdk_cx::SdkCallCx;
 pub use services::Services;
-pub use trigger_event::{DeadLetterEventDetail, KvEventOp, TriggerEvent};
-pub use validator::{ScriptValidator, ValidationError};
+pub use trigger_event::{
+    DeadLetterEventDetail, DocsEventOp, FilesEventOp, KvEventOp, TriggerEvent,
+};
+pub use validator::{ScriptValidator, ValidatedScript, ValidationError};
 pub use version::{API_VERSION, PRODUCT_VERSION, SDK_VERSION, WIRE_VERSION};

crates/shared/src/modules.rs

@@ -0,0 +1,75 @@
+//! `ModuleSource` — the v1.1.3 Rhai module-loading contract.
+//!
+//! The executor-core `PicloudModuleResolver` calls into this trait to
+//! load `kind = 'module'` scripts referenced by `import "<name>" as <alias>;`
+//! statements. The Postgres impl in `manager-core` reads from the
+//! `scripts` table; tests pin in-memory fakes.
+//!
+//! Implementations MUST derive `app_id` from `cx.app_id` and pass it
+//! to every backend query. The `name` argument carries only the
+//! script's name (the literal between the import quotes); the trait
+//! has no way to express a cross-app lookup. That asymmetry is the
+//! load-bearing cross-app isolation boundary — see `docs/sdk-shape.md`.
+
+use async_trait::async_trait;
+use chrono::{DateTime, Utc};
+use serde::{Deserialize, Serialize};
+use thiserror::Error;
+
+use crate::{AppId, ScriptId, SdkCallCx};
+
+/// A module script as returned by `ModuleSource::lookup`. Carries only
+/// the fields the resolver needs: the id (for diagnostics), the source
+/// (to compile), and `updated_at` (the cache-staleness comparator).
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct ModuleScript {
+    pub script_id: ScriptId,
+    pub app_id: AppId,
+    pub name: String,
+    pub source: String,
+    pub updated_at: DateTime<Utc>,
+}
+
+/// Lookup contract used by `PicloudModuleResolver`. `lookup` MUST
+/// scope by `cx.app_id`; cross-app reads must be unreachable.
+#[async_trait]
+pub trait ModuleSource: Send + Sync {
+    /// Resolve a module script by `(cx.app_id, name)`. Returns `None`
+    /// when no row exists, or when a row exists but its `kind` is
+    /// `'endpoint'` (endpoints are never importable). The resolver
+    /// surfaces `None` as `ErrorModuleNotFound` to Rhai.
+    async fn lookup(
+        &self,
+        cx: &SdkCallCx,
+        name: &str,
+    ) -> Result<Option<ModuleScript>, ModuleSourceError>;
+}
+
+/// Failure modes surfaced from `ModuleSource::lookup`. "Not found" is
+/// not exceptional — it's `Ok(None)`.
+#[derive(Debug, Error)]
+pub enum ModuleSourceError {
+    /// Backend (Postgres, network, etc.) unavailable or returned an
+    /// error. The string is safe to surface to a script (Rhai wraps
+    /// it in `ErrorModuleNotFound` with the module name + reason).
+    #[error("module backend error: {0}")]
+    Backend(String),
+}
+
+/// Stub used by the executor-core test harness so engine integration
+/// tests don't need a real DB-backed source. Every lookup returns
+/// `Ok(None)` — `import "x"` always errors as "module not found"
+/// under this impl.
+#[derive(Debug, Default, Clone, Copy)]
+pub struct NoopModuleSource;
+
+#[async_trait]
+impl ModuleSource for NoopModuleSource {
+    async fn lookup(
+        &self,
+        _cx: &SdkCallCx,
+        _name: &str,
+    ) -> Result<Option<ModuleScript>, ModuleSourceError> {
+        Ok(None)
+    }
+}

crates/shared/src/pubsub.rs

@@ -0,0 +1,161 @@
+//! `PubsubService` — the v1.1.5 durable pub/sub contract.
+//!
+//! `pubsub::publish_durable(topic, message)` writes to the universal
+//! outbox; the publish-time fan-out inserts one delivery row per
+//! matching `pubsub` trigger, and each delivery retries / dead-letters
+//! independently (the dispatcher already handles one-row-equals-one-
+//! dispatch — no dispatcher changes for pub/sub).
+//!
+//! `publish_ephemeral` is committed as a v1.2 addition — the suffix
+//! naming exists now so users learn "durable by default" from day one.
+//!
+//! Topic pattern matching runs in Rust (not SQL) so the trigger-select
+//! query stays simple. The matcher + validator live here in
+//! `picloud-shared` so the manager-core publish path, the admin trigger
+//! endpoint, and tests all agree on the rules.
+
+use async_trait::async_trait;
+use thiserror::Error;
+
+use crate::SdkCallCx;
+
+#[async_trait]
+pub trait PubsubService: Send + Sync {
+    /// Durable publish: writes the message to the outbox, fanned out to
+    /// every matching enabled `pubsub` trigger in `cx.app_id`. Succeeds
+    /// silently (zero rows written) when no trigger matches the topic.
+    async fn publish_durable(
+        &self,
+        cx: &SdkCallCx,
+        topic: &str,
+        message: serde_json::Value,
+    ) -> Result<(), PubsubError>;
+}
+
+#[derive(Debug, Error)]
+pub enum PubsubError {
+    /// Empty topic; rejected at the SDK boundary.
+    #[error("topic must not be empty")]
+    EmptyTopic,
+
+    /// Caller principal lacked the required capability. Only raised when
+    /// `cx.principal.is_some()` (script-as-gate; public HTTP skips it).
+    #[error("forbidden")]
+    Forbidden,
+
+    /// Serialization / validation failure on the message.
+    #[error("pubsub rejected: {0}")]
+    Rejected(String),
+
+    /// Anything else — Postgres unavailable, etc.
+    #[error("pubsub backend error: {0}")]
+    Unavailable(String),
+}
+
+/// Match a stored `topic_pattern` against a published `topic`.
+///
+/// - `"*"` matches every topic.
+/// - `"<prefix>.*"` matches any topic starting with `"<prefix>."`.
+/// - anything else is an exact match.
+///
+/// Mid-pattern wildcards (`*.created`, `a.*.b`) are NOT supported — they
+/// are rejected at trigger creation by [`validate_topic_pattern`], so
+/// the only patterns reaching this matcher are exact / prefix / `*`.
+#[must_use]
+pub fn topic_matches(pattern: &str, topic: &str) -> bool {
+    if pattern == "*" {
+        return true;
+    }
+    if let Some(prefix) = pattern.strip_suffix('*') {
+        // `prefix` retains the trailing '.', e.g. "user." for "user.*".
+        return topic.starts_with(prefix);
+    }
+    pattern == topic
+}
+
+/// Validate a subscription topic pattern. Accepts exactly: `"*"`
+/// (universal), `"<prefix>.*"` (prefix wildcard, single trailing star),
+/// or a literal with no `*` (exact). Everything else — mid-pattern
+/// wildcards, multiple stars, a star not at the end — is rejected.
+///
+/// # Errors
+///
+/// Returns `Err(message)` with `"unsupported pubsub topic pattern: …"`
+/// for any unsupported shape (or an empty pattern).
+pub fn validate_topic_pattern(pattern: &str) -> Result<(), String> {
+    if pattern.is_empty() {
+        return Err("unsupported pubsub topic pattern: <empty>".to_string());
+    }
+    if pattern == "*" {
+        return Ok(());
+    }
+    let stars = pattern.matches('*').count();
+    if stars == 0 {
+        return Ok(()); // exact
+    }
+    if stars == 1 && pattern.ends_with(".*") {
+        return Ok(()); // prefix wildcard
+    }
+    Err(format!("unsupported pubsub topic pattern: {pattern}"))
+}
+
+/// Stub for the test harness so executor-core integration tests can
+/// build a `Services` bundle without a database. Every call errors.
+#[derive(Debug, Default, Clone, Copy)]
+pub struct NoopPubsubService;
+
+#[async_trait]
+impl PubsubService for NoopPubsubService {
+    async fn publish_durable(
+        &self,
+        _cx: &SdkCallCx,
+        _topic: &str,
+        _message: serde_json::Value,
+    ) -> Result<(), PubsubError> {
+        Err(PubsubError::Unavailable("pubsub is not wired in".into()))
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn exact_match() {
+        assert!(topic_matches("user.created", "user.created"));
+        assert!(!topic_matches("user.created", "user.deleted"));
+        assert!(!topic_matches("user.created", "user.created.x"));
+    }
+
+    #[test]
+    fn prefix_wildcard() {
+        assert!(topic_matches("user.*", "user.created"));
+        assert!(topic_matches("user.*", "user.deleted"));
+        assert!(!topic_matches("user.*", "users.created"));
+        assert!(!topic_matches("user.*", "order.created"));
+    }
+
+    #[test]
+    fn universal() {
+        assert!(topic_matches("*", "anything"));
+        assert!(topic_matches("*", "a.b.c"));
+    }
+
+    #[test]
+    fn validation_accepts_supported_shapes() {
+        assert!(validate_topic_pattern("*").is_ok());
+        assert!(validate_topic_pattern("user.created").is_ok());
+        assert!(validate_topic_pattern("user.*").is_ok());
+        assert!(validate_topic_pattern("a.b.c").is_ok());
+    }
+
+    #[test]
+    fn validation_rejects_unsupported_shapes() {
+        for bad in ["*.created", "**", "a.*.b", "user.*x", "*user", ""] {
+            assert!(
+                validate_topic_pattern(bad).is_err(),
+                "expected {bad:?} to be rejected"
+            );
+        }
+    }
+}

crates/shared/src/script.rs

@@ -3,6 +3,89 @@ use serde::{Deserialize, Serialize};
 
 use crate::{AppId, ScriptId, ScriptSandbox};
 
+/// Semantic role of a script (v1.1.3).
+///
+/// `Endpoint` scripts have an executable entry point — they bind to HTTP
+/// routes and act as trigger handlers. `Module` scripts are libraries of
+/// `fn`/`const` declarations imported by other scripts via Rhai's
+/// `import "<name>" as <alias>;` syntax. Modules cannot be invoked
+/// directly: route binding and trigger creation reject `Module` targets.
+///
+/// Serialized as `"endpoint"` / `"module"` so the wire shape is the
+/// same string the SQL `CHECK (kind IN ('endpoint','module'))`
+/// constraint enforces.
+#[derive(Debug, Clone, Copy, Default, PartialEq, Eq, Hash, Serialize, Deserialize)]
+#[serde(rename_all = "lowercase")]
+pub enum ScriptKind {
+    #[default]
+    Endpoint,
+    Module,
+}
+
+impl ScriptKind {
+    /// Wire / SQL representation. Inverse of `parse_str`.
+    #[must_use]
+    pub fn as_str(self) -> &'static str {
+        match self {
+            Self::Endpoint => "endpoint",
+            Self::Module => "module",
+        }
+    }
+
+    /// Parse the canonical wire / SQL form. Returns `None` for any
+    /// other input; callers map that to a 400 / `ValidationError`.
+    /// Named `parse_str` (not `from_str`) to dodge the
+    /// `std::str::FromStr` lint without taking on the trait's
+    /// `Result<Self, Self::Err>` shape that this caller doesn't need.
+    #[must_use]
+    pub fn parse_str(s: &str) -> Option<Self> {
+        match s {
+            "endpoint" => Some(Self::Endpoint),
+            "module" => Some(Self::Module),
+            _ => None,
+        }
+    }
+}
+
+#[cfg(test)]
+mod kind_tests {
+    use super::*;
+
+    #[test]
+    fn default_is_endpoint() {
+        assert_eq!(ScriptKind::default(), ScriptKind::Endpoint);
+    }
+
+    #[test]
+    fn round_trips_through_serde_lowercase() {
+        assert_eq!(
+            serde_json::to_string(&ScriptKind::Endpoint).unwrap(),
+            "\"endpoint\""
+        );
+        assert_eq!(
+            serde_json::to_string(&ScriptKind::Module).unwrap(),
+            "\"module\""
+        );
+        assert_eq!(
+            serde_json::from_str::<ScriptKind>("\"endpoint\"").unwrap(),
+            ScriptKind::Endpoint
+        );
+        assert_eq!(
+            serde_json::from_str::<ScriptKind>("\"module\"").unwrap(),
+            ScriptKind::Module
+        );
+    }
+
+    #[test]
+    fn parse_str_round_trip() {
+        for k in [ScriptKind::Endpoint, ScriptKind::Module] {
+            assert_eq!(ScriptKind::parse_str(k.as_str()), Some(k));
+        }
+        assert_eq!(ScriptKind::parse_str("invalid"), None);
+        assert_eq!(ScriptKind::parse_str(""), None);
+    }
+}
+
 /// A user-uploaded Rhai script and its execution configuration.
 ///
 /// This is the canonical representation that flows between manager (storage),
@@ -20,6 +103,12 @@ pub struct Script {
     pub version: i32,
     pub source: String,
 
+    /// `Endpoint` (default; the only kind v1.0 through v1.1.2 supported)
+    /// or `Module` (v1.1.3 — imported by other scripts, never bound
+    /// directly to a route or trigger).
+    #[serde(default)]
+    pub kind: ScriptKind,
+
     pub timeout_seconds: u32,
 
     /// Per-script overrides for Rhai sandbox limits. Empty = platform

crates/shared/src/sdk_cx.rs

@@ -12,7 +12,7 @@
 //! the cx in is shared by both sides. Pure value type — no handles, no
 //! DB pool references, no allocations beyond what's in `Principal`.
 
-use crate::{AppId, ExecutionId, Principal, RequestId, TriggerEvent};
+use crate::{AppId, ExecutionId, Principal, RequestId, ScriptId, TriggerEvent};
 
 /// Per-invocation context for every stateful SDK service call.
 ///
@@ -27,6 +27,11 @@ pub struct SdkCallCx {
     /// every `(app_id, …)` storage lookup the script makes.
     pub app_id: AppId,
 
+    /// The script being executed. Used for audit-log attribution and
+    /// the default outbound-HTTP `User-Agent` (`picloud/<v>
+    /// (script:<id>)`). Added in v1.1.4 for the `http::*` SDK.
+    pub script_id: ScriptId,
+
     /// Caller identity, when authenticated. `None` for unauthenticated
     /// data-plane HTTP requests (the common case for public endpoints);
     /// `Some` when the call came in via the dashboard, an API key, or a

crates/shared/src/services.rs

@@ -7,9 +7,11 @@
 //! handed to `executor-core::sdk::register_all` alongside an
 //! `SdkCallCx` to wire each `::` namespace.
 //!
-//! v1.1.0 shipped this empty; v1.1.1 adds the first two service fields
+//! v1.1.0 shipped this empty; v1.1.1 added the first two service fields
 //! (`kv`, `dead_letters`) plus the `events` emitter that bound services
-//! use to publish events into the triggers outbox.
+//! use to publish events into the triggers outbox. v1.1.3 adds the
+//! `modules` field — the `ModuleSource` consulted by the per-call
+//! `PicloudModuleResolver` to load `import`ed module scripts.
 //!
 //! `#[non_exhaustive]` so adding fields is a non-breaking change for
 //! consumers that only *pattern-match* a `&Services`; only crates that
@@ -18,8 +20,9 @@
 use std::sync::Arc;
 
 use crate::{
-    DeadLetterService, KvService, NoopDeadLetterService, NoopEventEmitter, NoopKvService,
-    ServiceEventEmitter,
+    DeadLetterService, DocsService, FilesService, HttpService, KvService, ModuleSource,
+    NoopDeadLetterService, NoopDocsService, NoopEventEmitter, NoopFilesService, NoopHttpService,
+    NoopKvService, NoopModuleSource, NoopPubsubService, PubsubService, ServiceEventEmitter,
 };
 
 /// SDK service bundle. See module docs for the lifecycle and the v1.1.x
@@ -30,16 +33,46 @@ pub struct Services {
     /// in-memory in tests.
     pub kv: Arc<dyn KvService>,
 
+    /// Document store (v1.1.2). Backed by Postgres in the picloud
+    /// binary; in-memory in tests.
+    pub docs: Arc<dyn DocsService>,
+
     /// Dead-letter management (v1.1.1). Scripts get
     /// `dead_letters::replay(id)` and `dead_letters::resolve(id, reason)`.
     pub dead_letters: Arc<dyn DeadLetterService>,
 
     /// Event emitter for the triggers outbox. Mutating service methods
-    /// (`KvService::set/delete`, future `docs::*`, `files::*`, etc.)
-    /// call `events.emit(cx, event)` after the write succeeds. The
-    /// outbox-backed impl in `manager-core::outbox_event_emitter`
-    /// replaces v1.1.0's `NoopEventEmitter`.
+    /// (`KvService::set/delete`, `DocsService::create/update/delete`,
+    /// future `files::*`, etc.) call `events.emit(cx, event)` after
+    /// the write succeeds. The outbox-backed impl in
+    /// `manager-core::outbox_event_emitter` replaces v1.1.0's
+    /// `NoopEventEmitter`.
     pub events: Arc<dyn ServiceEventEmitter>,
+
+    /// Module source (v1.1.3). The `PicloudModuleResolver` consults
+    /// this to load `kind = 'module'` scripts that other scripts
+    /// `import`. Backed by Postgres in the picloud binary; in-memory
+    /// fakes in resolver tests.
+    pub modules: Arc<dyn ModuleSource>,
+
+    /// Outbound HTTP (v1.1.4). Scripts get `http::{get,post,…}`.
+    /// Backed by a reqwest client with the SSRF deny-list resolver in
+    /// the picloud binary; `NoopHttpService` in tests that don't make
+    /// network calls.
+    pub http: Arc<dyn HttpService>,
+
+    /// Filesystem-backed blob storage (v1.1.5). Scripts get
+    /// `files::collection(name).{create,head,get,update,delete,list}`.
+    /// Backed by a Postgres-metadata + on-disk-bytes repo in the
+    /// picloud binary; `NoopFilesService` in tests that don't touch
+    /// files.
+    pub files: Arc<dyn FilesService>,
+
+    /// Durable pub/sub (v1.1.5). Scripts get
+    /// `pubsub::publish_durable(topic, message)`. Backed by a
+    /// publish-time outbox fan-out in the picloud binary;
+    /// `NoopPubsubService` in tests that don't publish.
+    pub pubsub: Arc<dyn PubsubService>,
 }
 
 impl Services {
@@ -47,15 +80,26 @@ impl Services {
     /// The picloud binary's `main` wires this up after the DB pool is
     /// open; tests build it from in-memory fakes.
     #[must_use]
+    #[allow(clippy::too_many_arguments)] // one Arc per stateful service; a builder would just move the noise
     pub fn new(
         kv: Arc<dyn KvService>,
+        docs: Arc<dyn DocsService>,
         dead_letters: Arc<dyn DeadLetterService>,
         events: Arc<dyn ServiceEventEmitter>,
+        modules: Arc<dyn ModuleSource>,
+        http: Arc<dyn HttpService>,
+        files: Arc<dyn FilesService>,
+        pubsub: Arc<dyn PubsubService>,
     ) -> Self {
         Self {
             kv,
+            docs,
             dead_letters,
             events,
+            modules,
+            http,
+            files,
+            pubsub,
         }
     }
 
@@ -68,8 +112,13 @@ impl Services {
     pub fn with_noop_services() -> Self {
         Self::new(
             Arc::new(NoopKvService),
+            Arc::new(NoopDocsService),
             Arc::new(NoopDeadLetterService),
             Arc::new(NoopEventEmitter),
+            Arc::new(NoopModuleSource),
+            Arc::new(NoopHttpService),
+            Arc::new(NoopFilesService),
+            Arc::new(NoopPubsubService),
         )
     }
 }

crates/shared/src/trigger_event.rs

@@ -45,6 +45,72 @@ impl KvEventOp {
     }
 }
 
+/// Operations a docs trigger can fire on. v1.1.2. Stored as a
+/// lowercase string in `docs_trigger_details.ops` (Postgres `text[]`).
+/// Distinct from `KvEventOp` because docs has CRUD verbs (`create`)
+/// instead of KV's set/upsert flavour (`insert`).
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, Serialize, Deserialize)]
+#[serde(rename_all = "lowercase")]
+pub enum DocsEventOp {
+    Create,
+    Update,
+    Delete,
+}
+
+impl DocsEventOp {
+    #[must_use]
+    pub const fn as_str(self) -> &'static str {
+        match self {
+            Self::Create => "create",
+            Self::Update => "update",
+            Self::Delete => "delete",
+        }
+    }
+
+    #[must_use]
+    pub fn from_wire(s: &str) -> Option<Self> {
+        match s {
+            "create" => Some(Self::Create),
+            "update" => Some(Self::Update),
+            "delete" => Some(Self::Delete),
+            _ => None,
+        }
+    }
+}
+
+/// Operations a files trigger can fire on. v1.1.5. Stored as a
+/// lowercase string in `files_trigger_details.ops` (Postgres `text[]`).
+/// CRUD verbs (`create`) mirror `DocsEventOp`, distinct from KV's
+/// set/upsert flavour.
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash, Serialize, Deserialize)]
+#[serde(rename_all = "lowercase")]
+pub enum FilesEventOp {
+    Create,
+    Update,
+    Delete,
+}
+
+impl FilesEventOp {
+    #[must_use]
+    pub const fn as_str(self) -> &'static str {
+        match self {
+            Self::Create => "create",
+            Self::Update => "update",
+            Self::Delete => "delete",
+        }
+    }
+
+    #[must_use]
+    pub fn from_wire(s: &str) -> Option<Self> {
+        match s {
+            "create" => Some(Self::Create),
+            "update" => Some(Self::Update),
+            "delete" => Some(Self::Delete),
+            _ => None,
+        }
+    }
+}
+
 /// Discriminated description of a triggering event. Lifted from the
 /// outbox row's payload at dispatch time. Each variant carries the
 /// fields the corresponding `ctx.event` shape exposes to the script.
@@ -61,6 +127,65 @@ pub enum TriggerEvent {
         value: Option<serde_json::Value>,
     },
 
+    /// A docs create / update / delete fired this handler. v1.1.2.
+    /// `data` is the current document state (absent on delete);
+    /// `prev_data` is the prior state (absent on create). For update
+    /// and delete handlers, `prev_data` is the load-bearing
+    /// change-data-capture surface (the repo reads the old row in the
+    /// same statement as the write).
+    Docs {
+        op: DocsEventOp,
+        collection: String,
+        /// UUID as string — Rhai sees it as a string.
+        id: String,
+        #[serde(default, skip_serializing_if = "Option::is_none")]
+        data: Option<serde_json::Value>,
+        #[serde(default, skip_serializing_if = "Option::is_none")]
+        prev_data: Option<serde_json::Value>,
+    },
+
+    /// A cron schedule fired this handler. v1.1.4. Carries the
+    /// schedule + timezone the trigger was configured with, the
+    /// canonical cron moment (`scheduled_at`, the instant the
+    /// expression *meant*), and when the scheduler actually enqueued
+    /// the fire (`fired_at`). Surfaced to scripts as `ctx.event.cron`.
+    Cron {
+        schedule: String,
+        timezone: String,
+        scheduled_at: DateTime<Utc>,
+        fired_at: DateTime<Utc>,
+    },
+
+    /// A files create / update / delete fired this handler. v1.1.5.
+    /// Carries the affected file's **metadata only** — never the blob
+    /// bytes (files are too big to ship through trigger payloads). A
+    /// handler that wants the bytes calls
+    /// `files::collection(c).get(id)` itself. `prev` is the prior
+    /// metadata for update (and the deleted-row metadata for delete);
+    /// absent on create. Surfaced to scripts as `ctx.event.files`.
+    Files {
+        op: FilesEventOp,
+        collection: String,
+        /// UUID as string — Rhai sees it as a string.
+        id: String,
+        name: String,
+        content_type: String,
+        size: u64,
+        /// Lowercase hex SHA-256.
+        checksum: String,
+        #[serde(default, skip_serializing_if = "Option::is_none")]
+        prev: Option<serde_json::Value>,
+    },
+
+    /// A durable pub/sub publish fired this handler. v1.1.5. Carries
+    /// the topic, the JSON-decoded message, and the publish instant.
+    /// Surfaced to scripts as `ctx.event.pubsub`.
+    Pubsub {
+        topic: String,
+        message: serde_json::Value,
+        published_at: DateTime<Utc>,
+    },
+
     /// A dead-letter row fired this handler. The original event is
     /// nested verbatim plus the dead-letter metadata the design notes
     /// §4 require.
@@ -84,6 +209,10 @@ impl TriggerEvent {
     pub const fn source(&self) -> &'static str {
         match self {
             Self::Kv { .. } => "kv",
+            Self::Docs { .. } => "docs",
+            Self::Cron { .. } => "cron",
+            Self::Files { .. } => "files",
+            Self::Pubsub { .. } => "pubsub",
             Self::DeadLetter { .. } => "dead_letter",
         }
     }

crates/shared/src/validator.rs

@@ -10,8 +10,39 @@ use thiserror::Error;
 pub enum ValidationError {
     #[error("invalid script source: {0}")]
     Syntax(String),
+
+    /// v1.1.3: source compiled but failed the module-shape gate
+    /// (top-level statements other than `fn` / `const` / `import`).
+    #[error("module syntax error: {0}")]
+    ModuleShape(String),
+}
+
+/// Output of a successful validate. v1.1.3 carries the list of literal
+/// `import "<name>"` paths the script declares — the manager writes
+/// these into the `script_imports` dep-graph table. Endpoints may also
+/// have imports; the field is populated unconditionally.
+#[derive(Debug, Clone, Default)]
+pub struct ValidatedScript {
+    /// Literal-path imports (in declaration order). Dynamic imports
+    /// `import some_var as y;` are not captured — the resolver still
+    /// honors them at runtime, but the dep graph only tracks names
+    /// known at compile time.
+    pub imports: Vec<String>,
 }
 
 pub trait ScriptValidator: Send + Sync {
-    fn validate(&self, source: &str) -> Result<(), ValidationError>;
+    /// Endpoint-shape validation: parse-only syntax check. Returns the
+    /// declared (literal) imports so the manager can populate the
+    /// dep-graph table on save.
+    fn validate(&self, source: &str) -> Result<ValidatedScript, ValidationError>;
+
+    /// Module-shape validation: parse + reject any top-level
+    /// statement that isn't `fn` / `const` / `import`. Default impl
+    /// rejects every module so non-engine validators stay simple
+    /// (tests / stubs don't need to know module rules).
+    fn validate_module(&self, _source: &str) -> Result<ValidatedScript, ValidationError> {
+        Err(ValidationError::ModuleShape(
+            "module validation not implemented by this validator".into(),
+        ))
+    }
 }

crates/shared/src/version.rs

@@ -22,7 +22,34 @@ pub const PRODUCT_VERSION: &str = env!("CARGO_PKG_VERSION");
 ///
 /// 1.2 additions (v1.1.1): `kv::collection(name).{get,set,has,delete,list}`,
 /// `dead_letters::{replay,resolve}`, `ctx.event` for triggered handlers.
-pub const SDK_VERSION: &str = "1.2";
+///
+/// 1.3 additions (v1.1.2):
+/// `docs::collection(name).{create,get,find,find_one,update,delete,list}`
+/// with the v1.1.2 query DSL subset; `ctx.event.docs` for docs-trigger
+/// handlers (carries `prev_data` change-data-capture for update/delete).
+///
+/// 1.4 additions (v1.1.3): `import "<name>" as <alias>;` for scripts
+/// whose corresponding module (`kind = 'module'`) lives in the same
+/// app. Cross-app imports are unreachable (the `name` argument carries
+/// no `app_id`). Modules expose `fn`/`const` declarations only;
+/// top-level statements are rejected at create-time.
+///
+/// 1.5 additions (v1.1.4): `http::{get,post,put,patch,delete,head,
+/// post_form,request}` for outbound HTTP from scripts (guarded by an
+/// SSRF deny-list on the resolved IP); `ctx.event.cron` for cron-trigger
+/// handlers (carries `schedule`, `timezone`, `scheduled_at`, `fired_at`).
+/// The `Services` bundle gains `http: Arc<dyn HttpService>`.
+///
+/// 1.6 additions (v1.1.5):
+/// `files::collection(name).{create,head,get,update,delete,list}` —
+/// filesystem-backed blob storage (blobs in/out; metadata maps;
+/// checksum-verified reads) with `ctx.event.files` for files-trigger
+/// handlers (metadata only, never the bytes); and
+/// `pubsub::publish_durable(topic, message)` — durable pub/sub with
+/// publish-time fan-out and `ctx.event.pubsub` for pubsub-trigger
+/// handlers. The `Services` bundle gains `files: Arc<dyn FilesService>`
+/// and `pubsub: Arc<dyn PubsubService>`.
+pub const SDK_VERSION: &str = "1.6";
 
 /// HTTP API major version. Appears in URL paths as `/api/v{N}/...`.
 /// Bump (new integer + new URL prefix) when the request/response

dashboard/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "picloud-dashboard",
-	"version": "0.7.0",
+	"version": "0.11.0",
 	"private": true,
 	"type": "module",
 	"scripts": {

dashboard/src/lib/api.ts

@@ -21,6 +21,8 @@ export interface ScriptSandbox {
 	max_expr_depth?: number;
 }
 
+export type ScriptKind = 'endpoint' | 'module';
+
 export interface Script {
 	id: string;
 	app_id: string;
@@ -28,6 +30,8 @@ export interface Script {
 	description: string | null;
 	version: number;
 	source: string;
+	/** v1.1.3 — 'endpoint' (default) handles routes/triggers; 'module' is imported by other scripts. */
+	kind: ScriptKind;
 	timeout_seconds: number;
 	memory_limit_mb: number;
 	sandbox: ScriptSandbox;
@@ -173,6 +177,8 @@ export interface CreateScriptInput {
 	name: string;
 	description?: string | null;
 	source: string;
+	/** Defaults to 'endpoint' server-side if omitted. v1.1.3. */
+	kind?: ScriptKind;
 	timeout_seconds?: number;
 	memory_limit_mb?: number;
 }
@@ -184,6 +190,8 @@ export interface UpdateScriptInput {
 	timeout_seconds?: number;
 	memory_limit_mb?: number;
 	sandbox?: ScriptSandbox;
+	/** v1.1.3 — endpoint→module rejected if routes/triggers reference the script. */
+	kind?: ScriptKind;
 }
 
 export interface DeadLetterRow {
@@ -203,6 +211,65 @@ export interface DeadLetterRow {
 	resolution: 'replayed' | 'ignored' | 'handled_by_script' | 'handler_failed' | null;
 }
 
+export type TriggerKind = 'kv' | 'docs' | 'dead_letter' | 'cron' | 'files' | 'pubsub';
+export type TriggerDispatchMode = 'sync' | 'async';
+
+/// Per-kind detail, tagged by `kind` to match the Rust serde shape.
+export type TriggerDetails =
+	| { kind: 'kv'; collection_glob: string; ops: string[] }
+	| { kind: 'docs'; collection_glob: string; ops: string[] }
+	| { kind: 'dead_letter'; source_filter?: string; trigger_id_filter?: string; script_id_filter?: string }
+	| { kind: 'cron'; schedule: string; timezone: string; last_fired_at?: string | null }
+	| { kind: 'files'; collection_glob: string; ops: string[] }
+	| { kind: 'pubsub'; topic_pattern: string };
+
+/// v1.1.5 file metadata as the admin files endpoint returns it.
+export interface FileMeta {
+	id: string;
+	collection: string;
+	name: string;
+	content_type: string;
+	size: number;
+	checksum: string;
+	created_at: string;
+	updated_at: string;
+}
+
+export interface Trigger {
+	id: string;
+	app_id: string;
+	script_id: string;
+	kind: TriggerKind;
+	enabled: boolean;
+	dispatch_mode: TriggerDispatchMode;
+	retry_max_attempts: number;
+	retry_backoff: 'exponential' | 'linear' | 'constant';
+	retry_base_ms: number;
+	registered_by_principal: string;
+	created_at: string;
+	updated_at: string;
+	details: TriggerDetails;
+}
+
+export interface CreateCronTriggerInput {
+	script_id: string;
+	schedule: string;
+	timezone: string;
+	dispatch_mode?: TriggerDispatchMode;
+	retry_max_attempts?: number;
+	retry_backoff?: 'exponential' | 'linear' | 'constant';
+	retry_base_ms?: number;
+}
+
+export interface CreatePubsubTriggerInput {
+	script_id: string;
+	topic_pattern: string;
+	dispatch_mode?: TriggerDispatchMode;
+	retry_max_attempts?: number;
+	retry_backoff?: 'exponential' | 'linear' | 'constant';
+	retry_base_ms?: number;
+}
+
 export interface ExecutionResult {
 	status: number;
 	headers: Record<string, string>;
@@ -564,6 +631,45 @@ export const api = {
 			)
 	},
 
+	triggers: {
+		list: (idOrSlug: string) =>
+			adminRequest<{ triggers: Trigger[] }>(
+				`/api/v1/admin/apps/${encodeURIComponent(idOrSlug)}/triggers`
+			),
+		createCron: (idOrSlug: string, input: CreateCronTriggerInput) =>
+			adminRequest<Trigger>(
+				`/api/v1/admin/apps/${encodeURIComponent(idOrSlug)}/triggers/cron`,
+				{ method: 'POST', body: JSON.stringify(input) }
+			),
+		createPubsub: (idOrSlug: string, input: CreatePubsubTriggerInput) =>
+			adminRequest<Trigger>(
+				`/api/v1/admin/apps/${encodeURIComponent(idOrSlug)}/triggers/pubsub`,
+				{ method: 'POST', body: JSON.stringify(input) }
+			),
+		remove: (idOrSlug: string, triggerId: string) =>
+			adminRequest<null>(
+				`/api/v1/admin/apps/${encodeURIComponent(idOrSlug)}/triggers/${triggerId}`,
+				{ method: 'DELETE' }
+			)
+	},
+
+	files: {
+		list: (idOrSlug: string, collection: string, opts: { cursor?: string; limit?: number } = {}) => {
+			const params = new URLSearchParams();
+			params.set('collection', collection);
+			if (opts.cursor) params.set('cursor', opts.cursor);
+			if (opts.limit !== undefined) params.set('limit', String(opts.limit));
+			return adminRequest<{ files: FileMeta[]; next_cursor: string | null }>(
+				`/api/v1/admin/apps/${encodeURIComponent(idOrSlug)}/files?${params.toString()}`
+			);
+		},
+		remove: (idOrSlug: string, collection: string, fileId: string) =>
+			adminRequest<null>(
+				`/api/v1/admin/apps/${encodeURIComponent(idOrSlug)}/files/${encodeURIComponent(collection)}/${fileId}`,
+				{ method: 'DELETE' }
+			)
+	},
+
 	execute: async (
 		id: string,
 		body: unknown,

dashboard/src/routes/apps/[slug]/+page.svelte

@@ -10,7 +10,8 @@
 		type AppDomain,
 		type AppMemberDto,
 		type AppRole,
-		type Script
+		type Script,
+		type Trigger
 	} from '$lib/api';
 	import CodeEditor from '$lib/CodeEditor.svelte';
 	import ConfirmModal from '$lib/ConfirmModal.svelte';
@@ -24,7 +25,26 @@
 	const SAMPLE_SOURCE =
 		'#{\n  statusCode: 200,\n  body: #{ ok: true, echo: ctx.request.body }\n}';
 
-	type Tab = 'scripts' | 'domains' | 'members' | 'settings';
+	type Tab = 'scripts' | 'domains' | 'members' | 'settings' | 'triggers';
+
+	// Common IANA timezones offered in the cron form dropdown. Not
+	// exhaustive — the backend validates any IANA name via chrono-tz.
+	const COMMON_TIMEZONES = [
+		'UTC',
+		'America/Los_Angeles',
+		'America/Denver',
+		'America/Chicago',
+		'America/New_York',
+		'America/Sao_Paulo',
+		'Europe/London',
+		'Europe/Berlin',
+		'Europe/Paris',
+		'Europe/Moscow',
+		'Asia/Kolkata',
+		'Asia/Shanghai',
+		'Asia/Tokyo',
+		'Australia/Sydney'
+	];
 
 	let slug = $derived(page.params.slug ?? '');
 	let app = $state<App | null>(null);
@@ -63,6 +83,10 @@
 	let createScriptName = $state('');
 	let createScriptDescription = $state('');
 	let createScriptSource = $state(SAMPLE_SOURCE);
+	// v1.1.3: endpoint (default — handles routes/triggers) vs module
+	// (library imported by other scripts). Modules cannot be bound to
+	// routes or used as trigger targets.
+	let createScriptKind = $state<'endpoint' | 'module'>('endpoint');
 	let creatingScript = $state(false);
 	let createScriptError = $state<string | null>(null);
 
@@ -87,6 +111,89 @@
 	let removingDomain = $state(false);
 	let removeDomainError = $state<string | null>(null);
 
+	// Triggers tab (v1.1.4 — cron triggers). Admin-gated, like Members.
+	let triggers = $state<Trigger[]>([]);
+	let createCronScriptId = $state('');
+	let createCronSchedule = $state('0 0 9 * * MON-FRI');
+	let createCronTimezone = $state('UTC');
+	let creatingCron = $state(false);
+	let createCronError = $state<string | null>(null);
+	// Pub/Sub triggers (v1.1.5).
+	let createPubsubScriptId = $state('');
+	let createPubsubTopic = $state('');
+	let creatingPubsub = $state(false);
+	let createPubsubError = $state<string | null>(null);
+	let triggerToRemove = $state<Trigger | null>(null);
+	let removingTrigger = $state(false);
+	// Endpoint scripts only — modules can't be trigger targets.
+	const endpointScripts = $derived(scripts.filter((s) => s.kind === 'endpoint'));
+
+	async function loadTriggers(idOrSlug: string) {
+		try {
+			const r = await api.triggers.list(idOrSlug);
+			triggers = r.triggers;
+		} catch {
+			triggers = [];
+		}
+	}
+
+	async function submitCreateCron(e: SubmitEvent) {
+		e.preventDefault();
+		if (!app) return;
+		creatingCron = true;
+		createCronError = null;
+		try {
+			await api.triggers.createCron(app.id, {
+				script_id: createCronScriptId,
+				schedule: createCronSchedule.trim(),
+				timezone: createCronTimezone
+			});
+			createCronScriptId = '';
+			await loadTriggers(app.id);
+		} catch (err) {
+			createCronError =
+				err instanceof ApiError ? err.message : err instanceof Error ? err.message : String(err);
+		} finally {
+			creatingCron = false;
+		}
+	}
+
+	async function submitCreatePubsub(e: SubmitEvent) {
+		e.preventDefault();
+		if (!app) return;
+		creatingPubsub = true;
+		createPubsubError = null;
+		try {
+			await api.triggers.createPubsub(app.id, {
+				script_id: createPubsubScriptId,
+				topic_pattern: createPubsubTopic.trim()
+			});
+			createPubsubScriptId = '';
+			createPubsubTopic = '';
+			await loadTriggers(app.id);
+		} catch (err) {
+			createPubsubError =
+				err instanceof ApiError ? err.message : err instanceof Error ? err.message : String(err);
+		} finally {
+			creatingPubsub = false;
+		}
+	}
+
+	async function confirmRemoveTrigger() {
+		if (!app || !triggerToRemove) return;
+		removingTrigger = true;
+		try {
+			await api.triggers.remove(app.id, triggerToRemove.id);
+			triggerToRemove = null;
+			await loadTriggers(app.id);
+		} catch (err) {
+			createCronError =
+				err instanceof ApiError ? err.message : err instanceof Error ? err.message : String(err);
+		} finally {
+			removingTrigger = false;
+		}
+	}
+
 	// Members tab
 	let eligibleUsers = $state<AdminDto[]>([]);
 	let eligibleLoadError = $state<string | null>(null);
@@ -127,7 +234,7 @@
 				loadDeadLetterCount(app.id)
 			];
 			if (canAdmin) {
-				loaders.push(loadMembers(app.id), loadEligibleUsers());
+				loaders.push(loadMembers(app.id), loadEligibleUsers(), loadTriggers(app.id));
 			}
 			await Promise.all(loaders);
 		} catch (e) {
@@ -201,12 +308,14 @@
 				app_id: app.id,
 				name: createScriptName.trim(),
 				description: createScriptDescription.trim() || null,
-				source: createScriptSource
+				source: createScriptSource,
+				kind: createScriptKind
 			});
 			showCreateScript = false;
 			createScriptName = '';
 			createScriptDescription = '';
 			createScriptSource = SAMPLE_SOURCE;
+			createScriptKind = 'endpoint';
 			await loadScripts(app.id);
 		} catch (e) {
 			createScriptError = e instanceof Error ? e.message : String(e);
@@ -392,7 +501,10 @@
 	// backend still 403s the underlying calls, but no point showing an
 	// empty tab.
 	$effect(() => {
-		if (!canAdmin && (activeTab === 'settings' || activeTab === 'members')) {
+		if (
+			!canAdmin &&
+			(activeTab === 'settings' || activeTab === 'members' || activeTab === 'triggers')
+		) {
 			activeTab = 'scripts';
 		}
 	});
@@ -434,11 +546,23 @@
 				class:active={activeTab === 'members'}
 				onclick={() => (activeTab = 'members')}>Members ({members.length})</button
 			>
+			<button
+				type="button"
+				class:active={activeTab === 'triggers'}
+				onclick={() => (activeTab = 'triggers')}>Triggers ({triggers.length})</button
+			>
 			<button
 				type="button"
 				class:active={activeTab === 'settings'}
 				onclick={() => (activeTab = 'settings')}>Settings</button
 			>
+			<a
+				class="tab-link"
+				href="{base}/apps/{slug}/files"
+				title="Files — browse and delete stored blobs by collection"
+			>
+				Files
+			</a>
 			<a
 				class="tab-link"
 				href="{base}/apps/{slug}/dead-letters"
@@ -473,6 +597,13 @@
 							<span>Name</span>
 							<input bind:value={createScriptName} required placeholder="echo" />
 						</label>
+						<label>
+							<span>Kind</span>
+							<select bind:value={createScriptKind}>
+								<option value="endpoint">Endpoint (handles HTTP / triggers)</option>
+								<option value="module">Module (imported by other scripts)</option>
+							</select>
+						</label>
 						<label>
 							<span>Description</span>
 							<input bind:value={createScriptDescription} placeholder="optional" />
@@ -482,6 +613,13 @@
 						<span>Source (Rhai)</span>
 						<CodeEditor bind:value={createScriptSource} language="rhai" minHeight="14rem" />
 					</label>
+					{#if createScriptKind === 'module'}
+						<p class="muted small">
+							Modules expose <code>fn</code> and <code>const</code> declarations to other
+							scripts via <code>import "name" as alias;</code>. They cannot be bound to
+							routes or used as trigger targets.
+						</p>
+					{/if}
 					{#if createScriptError}
 						<div class="error">{createScriptError}</div>
 					{/if}
@@ -503,6 +641,11 @@
 								<div class="primary">
 									<strong>{script.name}</strong>
 									<span class="muted">v{script.version}</span>
+									{#if script.kind === 'module'}
+										<span class="kind-badge kind-module" title="Library imported by other scripts">module</span>
+									{:else}
+										<span class="kind-badge kind-endpoint" title="Handles HTTP routes and trigger events">endpoint</span>
+									{/if}
 								</div>
 								<div class="secondary muted">{script.description ?? '—'}</div>
 							</a>
@@ -673,6 +816,129 @@
 				</div>
 			{/if}
 		</section>
+	{:else if activeTab === 'triggers' && canAdmin}
+		<section>
+			<h2>Cron triggers</h2>
+			<p class="muted">
+				Run an endpoint script on a schedule. Schedules are 6-field cron
+				expressions (with seconds): <code>sec min hour day-of-month month day-of-week</code>.
+				The timezone disambiguates schedules like "every weekday at 9am".
+			</p>
+
+			<form class="create-form" onsubmit={submitCreateCron}>
+				<div class="row">
+					<label>
+						<span>Target script</span>
+						<select bind:value={createCronScriptId} required>
+							<option value="" disabled>Select an endpoint script…</option>
+							{#each endpointScripts as s (s.id)}
+								<option value={s.id}>{s.name}</option>
+							{/each}
+						</select>
+					</label>
+					<label>
+						<span>Schedule</span>
+						<input
+							bind:value={createCronSchedule}
+							required
+							placeholder="0 0 9 * * MON-FRI"
+						/>
+					</label>
+					<label>
+						<span>Timezone</span>
+						<select bind:value={createCronTimezone}>
+							{#each COMMON_TIMEZONES as tz (tz)}
+								<option value={tz}>{tz}</option>
+							{/each}
+						</select>
+					</label>
+				</div>
+				{#if endpointScripts.length === 0}
+					<p class="muted small">
+						This app has no endpoint scripts yet — create one first (modules
+						can't be trigger targets).
+					</p>
+				{/if}
+				{#if createCronError}
+					<div class="error">{createCronError}</div>
+				{/if}
+				<div class="actions">
+					<button type="submit" disabled={creatingCron || !createCronScriptId}>
+						{creatingCron ? 'Creating…' : 'Create cron trigger'}
+					</button>
+				</div>
+			</form>
+
+			<h2>Pub/Sub triggers</h2>
+			<p class="muted">
+				Subscribe an endpoint script to durable pub/sub messages. Topic
+				patterns are an exact topic (<code>user.created</code>), a prefix
+				wildcard (<code>user.*</code>), or <code>*</code> for every topic.
+			</p>
+
+			<form class="create-form" onsubmit={submitCreatePubsub}>
+				<div class="row">
+					<label>
+						<span>Target script</span>
+						<select bind:value={createPubsubScriptId} required>
+							<option value="" disabled>Select an endpoint script…</option>
+							{#each endpointScripts as s (s.id)}
+								<option value={s.id}>{s.name}</option>
+							{/each}
+						</select>
+					</label>
+					<label>
+						<span>Topic pattern</span>
+						<input bind:value={createPubsubTopic} required placeholder="user.*" />
+					</label>
+				</div>
+				{#if createPubsubError}
+					<div class="error">{createPubsubError}</div>
+				{/if}
+				<div class="actions">
+					<button
+						type="submit"
+						disabled={creatingPubsub || !createPubsubScriptId || !createPubsubTopic.trim()}
+					>
+						{creatingPubsub ? 'Creating…' : 'Create pub/sub trigger'}
+					</button>
+				</div>
+			</form>
+
+			{#if triggers.length === 0}
+				<p class="muted">No triggers in this app yet.</p>
+			{:else}
+				<ul class="list">
+					{#each triggers as t (t.id)}
+						<li class="domain-row">
+							<div>
+								<span class="kind-badge">{t.kind}</span>
+								{#if t.details.kind === 'cron'}
+									<code>{t.details.schedule}</code>
+									<span class="muted">— {t.details.timezone}</span>
+									<span class="muted small">
+										last fired: {t.details.last_fired_at ?? 'never'}
+									</span>
+								{:else if t.details.kind === 'kv' || t.details.kind === 'docs' || t.details.kind === 'files'}
+									<code>{t.details.collection_glob}</code>
+									<span class="muted">— {t.details.ops.join(', ') || 'any op'}</span>
+								{:else if t.details.kind === 'pubsub'}
+									<code>{t.details.topic_pattern}</code>
+								{/if}
+								<span class="muted small">→ {t.script_id}</span>
+							</div>
+							<button
+								type="button"
+								class="secondary danger"
+								onclick={() => (triggerToRemove = t)}
+							>
+								Delete
+							</button>
+						</li>
+					{/each}
+				</ul>
+			{/if}
+		</section>
 	{:else if activeTab === 'settings' && canAdmin}
 		<section>
 			<h2>Settings</h2>
@@ -830,6 +1096,23 @@
 			{/if}
 		</ConfirmModal>
 	{/if}
+
+	{#if triggerToRemove}
+		<ConfirmModal
+			title="Delete trigger"
+			variant="danger"
+			confirmLabel="Delete trigger"
+			busyLabel="Deleting…"
+			busy={removingTrigger}
+			onConfirm={confirmRemoveTrigger}
+			onCancel={() => (triggerToRemove = null)}
+		>
+			<p>
+				This {triggerToRemove.kind} trigger will stop firing. The target
+				script is not affected.
+			</p>
+		</ConfirmModal>
+	{/if}
 {/if}
 
 <style>
@@ -1154,4 +1437,30 @@
 		display: flex;
 		justify-content: flex-end;
 	}
+
+	.kind-badge {
+		display: inline-block;
+		padding: 0 0.45rem;
+		margin-left: 0.5rem;
+		border-radius: 0.25rem;
+		font-size: 0.7rem;
+		font-weight: 600;
+		text-transform: uppercase;
+		letter-spacing: 0.05em;
+		line-height: 1.5;
+	}
+
+	.kind-endpoint {
+		background: #1e3a5f;
+		color: #93c5fd;
+	}
+
+	.kind-module {
+		background: #3f2e7d;
+		color: #c4b5fd;
+	}
+
+	.small {
+		font-size: 0.85rem;
+	}
 </style>

dashboard/src/routes/apps/[slug]/files/+page.svelte

@@ -0,0 +1,229 @@
+<script lang="ts">
+	import { base } from '$app/paths';
+	import { page } from '$app/state';
+	import { api, ApiError, type App, type FileMeta } from '$lib/api';
+	import ConfirmModal from '$lib/ConfirmModal.svelte';
+
+	let slug = $derived(page.params.slug ?? '');
+	let app = $state<App | null>(null);
+	let collection = $state('');
+	let activeCollection = $state('');
+	let files = $state<FileMeta[]>([]);
+	let nextCursor = $state<string | null>(null);
+	let loading = $state(false);
+	let error = $state<string | null>(null);
+	let fileToRemove = $state<FileMeta | null>(null);
+	let removing = $state(false);
+
+	async function loadApp() {
+		try {
+			app = await api.apps.get(slug);
+		} catch (e) {
+			error = e instanceof ApiError ? e.message : String(e);
+		}
+	}
+
+	$effect(() => {
+		void slug;
+		void loadApp();
+	});
+
+	async function loadFiles(cursor?: string) {
+		const c = collection.trim();
+		if (!c) {
+			error = 'Enter a collection name to list its files.';
+			return;
+		}
+		loading = true;
+		error = null;
+		try {
+			const res = await api.files.list(slug, c, { cursor, limit: 100 });
+			if (cursor) {
+				files = [...files, ...res.files];
+			} else {
+				files = res.files;
+				activeCollection = c;
+			}
+			nextCursor = res.next_cursor;
+		} catch (e) {
+			error = e instanceof ApiError ? e.message : String(e);
+		} finally {
+			loading = false;
+		}
+	}
+
+	async function confirmRemove() {
+		if (!fileToRemove) return;
+		removing = true;
+		try {
+			await api.files.remove(slug, fileToRemove.collection, fileToRemove.id);
+			files = files.filter((f) => f.id !== fileToRemove!.id);
+			fileToRemove = null;
+		} catch (e) {
+			error = e instanceof ApiError ? e.message : String(e);
+		} finally {
+			removing = false;
+		}
+	}
+
+	function fmtTime(iso: string): string {
+		return new Date(iso).toLocaleString();
+	}
+
+	function fmtSize(bytes: number): string {
+		if (bytes < 1024) return `${bytes} B`;
+		if (bytes < 1024 * 1024) return `${(bytes / 1024).toFixed(1)} KB`;
+		return `${(bytes / (1024 * 1024)).toFixed(1)} MB`;
+	}
+</script>
+
+<svelte:head>
+	<title>Files · {slug} · PiCloud</title>
+</svelte:head>
+
+<div class="container">
+	<header>
+		<div>
+			<a href="{base}/apps/{slug}" class="back">&larr; back to {app?.name ?? slug}</a>
+			<h1>Files</h1>
+			<p class="subtitle">
+				Browse and delete stored blobs by collection. Uploads happen from scripts via
+				<code>files::collection(c).create(…)</code>.
+			</p>
+		</div>
+	</header>
+
+	<form
+		class="collection-form"
+		onsubmit={(e) => {
+			e.preventDefault();
+			void loadFiles();
+		}}
+	>
+		<label>
+			<span>Collection</span>
+			<input bind:value={collection} placeholder="avatars" required />
+		</label>
+		<button type="submit" disabled={loading || !collection.trim()}>
+			{loading ? 'Loading…' : 'List files'}
+		</button>
+	</form>
+
+	{#if error}
+		<div class="error">{error}</div>
+	{/if}
+
+	{#if activeCollection}
+		{#if files.length === 0 && !loading}
+			<p class="muted">No files in collection <code>{activeCollection}</code>.</p>
+		{:else}
+			<table>
+				<thead>
+					<tr>
+						<th>Name</th>
+						<th>Content type</th>
+						<th>Size</th>
+						<th>Created</th>
+						<th>ID</th>
+						<th></th>
+					</tr>
+				</thead>
+				<tbody>
+					{#each files as f (f.id)}
+						<tr>
+							<td>{f.name}</td>
+							<td><code>{f.content_type}</code></td>
+							<td>{fmtSize(f.size)}</td>
+							<td>{fmtTime(f.created_at)}</td>
+							<td class="mono small">{f.id}</td>
+							<td>
+								<button type="button" class="danger" onclick={() => (fileToRemove = f)}>
+									Delete
+								</button>
+							</td>
+						</tr>
+					{/each}
+				</tbody>
+			</table>
+			{#if nextCursor}
+				<button type="button" class="secondary" onclick={() => loadFiles(nextCursor ?? undefined)}>
+					Load more
+				</button>
+			{/if}
+		{/if}
+	{/if}
+</div>
+
+{#if fileToRemove}
+	<ConfirmModal
+		title="Delete file"
+		variant="danger"
+		confirmLabel="Delete file"
+		onConfirm={confirmRemove}
+		onCancel={() => (fileToRemove = null)}
+	>
+		<p>
+			Delete <strong>{fileToRemove.name}</strong> ({fmtSize(fileToRemove.size)}) from collection
+			<code>{fileToRemove.collection}</code>? This removes both the metadata row and the bytes on
+			disk and cannot be undone.
+		</p>
+		{#if removing}<p class="muted">Deleting…</p>{/if}
+	</ConfirmModal>
+{/if}
+
+<style>
+	.container {
+		max-width: 60rem;
+		margin: 0 auto;
+		padding: 1.5rem;
+	}
+	header {
+		margin-bottom: 1rem;
+	}
+	.back {
+		font-size: 0.85rem;
+	}
+	.subtitle {
+		color: var(--muted, #666);
+		font-size: 0.9rem;
+	}
+	.collection-form {
+		display: flex;
+		align-items: flex-end;
+		gap: 0.75rem;
+		margin-bottom: 1rem;
+	}
+	.collection-form label {
+		display: flex;
+		flex-direction: column;
+		gap: 0.25rem;
+		font-size: 0.85rem;
+	}
+	table {
+		width: 100%;
+		border-collapse: collapse;
+		font-size: 0.9rem;
+	}
+	th,
+	td {
+		text-align: left;
+		padding: 0.4rem 0.6rem;
+		border-bottom: 1px solid var(--border, #e2e2e2);
+	}
+	.mono {
+		font-family: monospace;
+	}
+	.small {
+		font-size: 0.78rem;
+	}
+	.muted {
+		color: var(--muted, #666);
+	}
+	.error {
+		color: #b00020;
+		margin: 0.5rem 0;
+	}
+	button.danger {
+		color: #b00020;
+	}
+</style>

dashboard/src/routes/scripts/[id]/+page.svelte

@@ -433,7 +433,14 @@
 						<code>{script.name}</code>
 					</div>
 				{/if}
-				<h1>{script.name}</h1>
+				<h1>
+					{script.name}
+					{#if script.kind === 'module'}
+						<span class="kind-badge kind-module" title="Library imported by other scripts">module</span>
+					{:else}
+						<span class="kind-badge kind-endpoint" title="Handles HTTP routes and trigger events">endpoint</span>
+					{/if}
+				</h1>
 				<p class="muted">
 					v{script.version} · timeout {script.timeout_seconds}s · {script.description ?? 'no description'}
 				</p>
@@ -1323,4 +1330,27 @@
 		margin: 0.25rem 0 0 4rem;
 		font-size: 0.75rem;
 	}
+
+	.kind-badge {
+		display: inline-block;
+		padding: 0 0.45rem;
+		margin-left: 0.5rem;
+		border-radius: 0.25rem;
+		font-size: 0.65rem;
+		font-weight: 600;
+		text-transform: uppercase;
+		letter-spacing: 0.05em;
+		line-height: 1.5;
+		vertical-align: middle;
+	}
+
+	.kind-endpoint {
+		background: #1e3a5f;
+		color: #93c5fd;
+	}
+
+	.kind-module {
+		background: #3f2e7d;
+		color: #c4b5fd;
+	}
 </style>

serverless_cloud_blueprint.md

@@ -1772,7 +1772,7 @@ if allowed {
 - [ ] **Debugging**: How to trace interceptor execution in logs/dashboard?
 
 ### Rhai & SDK
-- [ ] **Module loading**: Can scripts `import` external Rhai modules? (probably no for MVP)
+- [x] **Module loading**: Scripts can `import "<name>" as <alias>;` other scripts in the same app (v1.1.3 — `scripts.kind = 'module'`). Per-app, cross-app isolated, cache-invalidated on `updated_at` change. External (off-platform) modules remain out of scope.
 - [ ] **File system access**: Can scripts read/write to local filesystem? (no for MVP)
 - [ ] **Request/response sizes**: Max payload size? (set sensible default, e.g., 10MB)