feat(safety): report, block and hide chat lines (MODULE_API 1.21.0)

USER_SAFETY.md step 3, the module half of report, block and hide.

- server/safety.js registers the rust.chat report target (snapshot for one
  reader, null when they could not have seen the line; staff's hide and
  unhide) and the steam game identity (ownersOf from rust_account_links).
- The events feed filters chat per reader, in the query: lines from a
  blocked account's Steam ids, the reader's unlinked Steam-ID blocks and the
  lines they hid. Signed-in readers get canBlock on each chat line (false for
  their own and for staff-linked players). Lines staff hid for everyone are
  left out for everyone but staff, who get them with hidden: true.
- rust_events gains hidden_at and hidden_by (ALTER ... IF NOT EXISTS).
- POST /admin/rust/chat/:id/hide and /unhide (admin, moderator), logged.
- The feed's chat lines get a menu: Report, Hide for me, Block player,
  and for staff Hide for everyone / Show to everyone again.
- coreApi ^1.21.0; ci/core-ref.json pinned to the website#233 merge;
  routes.manifest.json and swagger-fragment.json regenerated.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E14m6SuuY6i1vASFeGDBeY

ci/core-ref.json

@@ -1,6 +1,6 @@
 {
-  "$comment": "The core this module is proved against. MODULE_API.md §5.3: the frozen-manifest job clones RunicGateway/website at this exact ref, drops this module in as modules/rust and runs CORE's own routeManifest.js — nothing else can answer whether the URLs the module claims are the URLs it actually serves, because a mount prefix is a string in server/index.js and a documented path is a string in a JSON file, and whether those name the same URL is a fact about a running core. It also answers the blind spot phase 1 had to check by hand: core mounts several routes at the TIER ROOT (/status, /version), which the loader's collision probe cannot see, so /rust being free is asserted here by a core rather than by a reading. Pinned rather than tracking a branch on purpose: core moves for reasons that have nothing to do with this module, and a bump is then a deliberate commit saying which core the module was last proved against, instead of an unexplained red X on someone else's PR. Bump it, regenerate routes.manifest.json, and commit both together. This module needs MODULE_API 1.10.0 (module.json's coreApi is ^1.10.0), which the Event System cutover put on `main` — so unlike Module-uo, which spent the Event System window pinned to `edge`, this repo starts pinned to `main` and should stay there unless it comes to depend on a contract member that has not shipped yet. **2026-09-27, protocol 13 step 2:** this module calls ctx.events.expired (MODULE_API 1.11.0, PLAN_FIXES D183), so coreApi is ^1.11.0. It was pinned to website#209's branch head while that PR was open — a 1.10.0 core refuses to load it — and is back on `main` at the sha #209 merged as. **2026-10-12, protocol 14 (PLAN_REDESIGNS §2):** this module registers an event template and declares param durations, a lease value example and option notes (MODULE_API 1.13.0), so coreApi is ^1.13.0. It was pinned to website#212's branch head while that PR was open, and is back on `main` at the sha #212 merged as (the same tree). **2026-10-12, PLAN_REDESIGNS §12 (permission rewards):** this module declares a param with multiple: true (MODULE_API 1.14.0), so coreApi is ^1.14.0. It was pinned to website#213's branch head while that PR was open, and is back on `main` at the sha #213 merged as (the same tree). **2026-10-15, PLAN_REDESIGNS §14.3.4 (Dangerous Treasures):** the Start a Dangerous Treasures event step answers DT's warm-up with retryAfterMs (MODULE_API 1.15.0, D371), so coreApi is ^1.15.0, pinned to `main` at the sha website#215 merged as. **2026-10-19, Steam sign-in (PLAN_REDESIGNS §15, D392/D396):** this module registers onIdentityLinking and onIdentityLinked (MODULE_API 1.17.0), so coreApi is ^1.17.0. Pinned to the head of website feat/steam-signin while that PR is open — a 1.16.0 core refuses to load it — and is now on `edge` at the sha website#218 merged as. That is `edge`, not `main`, because #218 went into `edge` for the in-progress cutover. **2026-10-22, the cutover (PLAN_LAST_FIXES stage 6):** website#220 put `edge` on `main`, so this is back on `main` at the sha #220 merged as. That tree adds the last fixes' core half (website#219, no MODULE_API bump), whose only route changes are core's own event routes, so routes.manifest.json does not change. **2026-10-26, the public demo (DEMO_MODE.md §3.4, §3.5, DM18):** this module declares demoLocks, reads ctx.demoMode, registers onDemoSeed (MODULE_API 1.18.0) and gives its two console-command actions demoRefuses (1.19.0), so coreApi is ^1.19.0. It was pinned to the head of website feature/demo-step-refusal (website#224) while that PR was open — a 1.18.0 core never asks demoRefuses — and is back on `main` at the sha #224 merged as (the same tree, so routes.manifest.json does not change). **2026-10-30, account deletion (ACCOUNT_DELETION.md §4.4):** this module registers onUserDeleting (MODULE_API 1.20.0), so coreApi is ^1.20.0, pinned to `main` at the sha website#232 merged as.",
+  "$comment": "The core this module is proved against. MODULE_API.md §5.3: the frozen-manifest job clones RunicGateway/website at this exact ref, drops this module in as modules/rust and runs CORE's own routeManifest.js — nothing else can answer whether the URLs the module claims are the URLs it actually serves, because a mount prefix is a string in server/index.js and a documented path is a string in a JSON file, and whether those name the same URL is a fact about a running core. It also answers the blind spot phase 1 had to check by hand: core mounts several routes at the TIER ROOT (/status, /version), which the loader's collision probe cannot see, so /rust being free is asserted here by a core rather than by a reading. Pinned rather than tracking a branch on purpose: core moves for reasons that have nothing to do with this module, and a bump is then a deliberate commit saying which core the module was last proved against, instead of an unexplained red X on someone else's PR. Bump it, regenerate routes.manifest.json, and commit both together. This module needs MODULE_API 1.10.0 (module.json's coreApi is ^1.10.0), which the Event System cutover put on `main` — so unlike Module-uo, which spent the Event System window pinned to `edge`, this repo starts pinned to `main` and should stay there unless it comes to depend on a contract member that has not shipped yet. **2026-09-27, protocol 13 step 2:** this module calls ctx.events.expired (MODULE_API 1.11.0, PLAN_FIXES D183), so coreApi is ^1.11.0. It was pinned to website#209's branch head while that PR was open — a 1.10.0 core refuses to load it — and is back on `main` at the sha #209 merged as. **2026-10-12, protocol 14 (PLAN_REDESIGNS §2):** this module registers an event template and declares param durations, a lease value example and option notes (MODULE_API 1.13.0), so coreApi is ^1.13.0. It was pinned to website#212's branch head while that PR was open, and is back on `main` at the sha #212 merged as (the same tree). **2026-10-12, PLAN_REDESIGNS §12 (permission rewards):** this module declares a param with multiple: true (MODULE_API 1.14.0), so coreApi is ^1.14.0. It was pinned to website#213's branch head while that PR was open, and is back on `main` at the sha #213 merged as (the same tree). **2026-10-15, PLAN_REDESIGNS §14.3.4 (Dangerous Treasures):** the Start a Dangerous Treasures event step answers DT's warm-up with retryAfterMs (MODULE_API 1.15.0, D371), so coreApi is ^1.15.0, pinned to `main` at the sha website#215 merged as. **2026-10-19, Steam sign-in (PLAN_REDESIGNS §15, D392/D396):** this module registers onIdentityLinking and onIdentityLinked (MODULE_API 1.17.0), so coreApi is ^1.17.0. Pinned to the head of website feat/steam-signin while that PR is open — a 1.16.0 core refuses to load it — and is now on `edge` at the sha website#218 merged as. That is `edge`, not `main`, because #218 went into `edge` for the in-progress cutover. **2026-10-22, the cutover (PLAN_LAST_FIXES stage 6):** website#220 put `edge` on `main`, so this is back on `main` at the sha #220 merged as. That tree adds the last fixes' core half (website#219, no MODULE_API bump), whose only route changes are core's own event routes, so routes.manifest.json does not change. **2026-10-26, the public demo (DEMO_MODE.md §3.4, §3.5, DM18):** this module declares demoLocks, reads ctx.demoMode, registers onDemoSeed (MODULE_API 1.18.0) and gives its two console-command actions demoRefuses (1.19.0), so coreApi is ^1.19.0. It was pinned to the head of website feature/demo-step-refusal (website#224) while that PR was open — a 1.18.0 core never asks demoRefuses — and is back on `main` at the sha #224 merged as (the same tree, so routes.manifest.json does not change). **2026-10-30, account deletion (ACCOUNT_DELETION.md §4.4):** this module registers onUserDeleting (MODULE_API 1.20.0), so coreApi is ^1.20.0, pinned to `main` at the sha website#232 merged as. **2026-10-30, report, block and hide (USER_SAFETY.md §4.2):** this module registers a rust.chat report target and the steam game identity, and reads ctx.safety (MODULE_API 1.21.0), so coreApi is ^1.21.0, pinned to `main` at the sha website#233 merged as.",
   "repo": "https://gitea.whitlocktech.com/RunicGateway/website.git",
-  "ref": "7bd93b8bb0da734a642587f256d10c523ca39bea",
-  "refName": "website main @ website#232 merge, MODULE_API 1.20.0"
+  "ref": "f13b8fe1fec9dc330a985430b8bd4f67c9ec26c5",
+  "refName": "website main @ website#233 merge, MODULE_API 1.21.0"
 }