fix(security): harden authz, media gating, recover, SSE, deps

Security audit follow-up. No Critical/High existed; these close the
Medium/Low findings.

F1 [Med] Host privilege boundary: a host could demote a peer host to
  guest and then ban / PIN-reset (→ account takeover via /recover) them,
  because the ban/pin-reset peer guards key off the target's *current*
  role. set_role now blocks a non-admin from demoting a host.

F2 [Med] Moderation now revokes preview/thumbnail access: they are served
  through visibility-checked aliases (/api/v1/upload/{id}/{preview,
  thumbnail}) that filter soft-deleted / ban-hidden uploads, and direct
  /media/previews|thumbnails is 404-blocked. Feed emits the gated URLs;
  Caddy keeps them privately cacheable (max-age=300, short so moderation
  revokes promptly — the live feed already evicts cards via SSE). Uses a
  lean find_visible_media() (paths+mime only) on the media hot path.

F3 [Med/Low] npm audit fix: svelte 5.55.1→5.56.4 (SSR-XSS advisories),
  devalue 5.6.4→5.8.1 (DoS). Prod deps: 0 vulnerabilities.

F4 [Low] /recover no longer leaks account existence: unknown display name
  returns the same 401 as a wrong PIN, and runs a dummy bcrypt verify so
  timing matches — closing the enumeration + timing oracle.

F5 [Low] SSE streams re-validate the session every ~60s and close once it
  is logged out / expired, instead of running until client disconnect.

F6 [Info] X-Content-Type-Options: nosniff set at the app layer on all
  media responses (defense-in-depth alongside the edge).

Tests: new e2e specs for F1 (host cannot demote peer host), F2 (preview
gated + 404 after delete + direct /media blocked), F4 (uniform 401).
40 backend unit tests + 182 e2e passing.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

backend/src/auth/handlers.rs

@@ -121,6 +121,18 @@ pub struct RecoverResponse {
     pub user_id: Uuid,
 }
 
+/// A real cost-12 bcrypt hash of a fixed dummy value, computed once and cached. Used
+/// to run a constant-time-ish verify on the "unknown display name" branch of
+/// [`recover`], so that branch costs the same as a real (user-exists) verify and can't
+/// be told apart by timing.
+fn dummy_pin_hash() -> &'static str {
+    static HASH: std::sync::OnceLock<String> = std::sync::OnceLock::new();
+    HASH.get_or_init(|| {
+        bcrypt::hash("eventsnap-dummy-not-a-real-pin", 12)
+            .expect("bcrypt hashing of a static dummy value cannot fail")
+    })
+}
+
 pub async fn recover(
     State(state): State<AppState>,
     headers: HeaderMap,
@@ -158,9 +170,13 @@ pub async fn recover(
         User::find_by_event_and_name(&state.pool, event.id, display_name).await?;
 
     if users.is_empty() {
-        return Err(AppError::NotFound(
-            "Kein Benutzer mit diesem Namen gefunden.".into(),
-        ));
+        // No user with this name. Run a throwaway bcrypt verify so this branch takes
+        // the same time as the user-exists path, and return the SAME error as a wrong
+        // PIN — so "no such name" and "wrong PIN" are indistinguishable by response or
+        // timing. Display names are already public on the feed, but this still closes
+        // the /recover enumeration + timing oracle.
+        let _ = bcrypt::verify(&body.pin, dummy_pin_hash());
+        return Err(AppError::Unauthorized("PIN ist falsch.".into()));
     }
 
     for user in &users {