feat: crawler scaffold with chromium launcher (0.22.0)

- crawler module (browser, source trait, jobs, diff) + binary
- chromiumoxide launcher with fetcher feature (auto-downloads
  Chromium on first run, caches under ~/.cache/mangalord/chromium)
- LaunchOptions struct with extra_args, parseable from
  CRAWLER_BROWSER_MODE and CRAWLER_BROWSER_ARGS
- migration 0012 introduces sources, manga_sources,
  chapter_sources, crawler_jobs
- integration tests for headed + headless launch, ipify load+parse,
  and extra-args propagation (all #[ignore], opt-in)

backend/src/crawler/mod.rs

@@ -0,0 +1,19 @@
+//! Crawler subsystem.
+//!
+//! Runs as its own binary (`src/bin/crawler.rs`) and shares `domain`,
+//! `repo`, and `storage` with the API binary. Layering mirrors the
+//! `Storage` trait pattern: callers depend on the `source::Source`
+//! trait, not on a concrete site; new sites plug in as additional
+//! impls without touching the job runner.
+//!
+//! Submodules:
+//! - [`browser`]: launches and pools Chromium via `chromiumoxide`.
+//!   First run downloads a known-good build via the `fetcher` feature.
+//! - [`source`]: the `Source` trait. Per-site impls live alongside it.
+//! - [`jobs`]: job kinds, queue wrapper, handler dispatch.
+//! - [`diff`]: change detection — new / updated / dropped semantics.
+
+pub mod browser;
+pub mod diff;
+pub mod jobs;
+pub mod source;