refactor: remove voice, recording, and music/media features (gateway + backend)

- Gateway: delete voice-recording/, voice-pcm-ws/, voice/video/media handlers,
  vendor/discord-voice-fork/, voice DB repos, 2 voice migrations
- Gateway: cut voice wiring from bootstrap/shutdown/commandHandler/handler-registry,
  eventBroadcaster/eventTypes, redis-channels, moderation-types, message-capture,
  config keys, and deps (@discordjs/voice, opus, libsodium, prism-media, davey)
- Backend: delete voice/, recordings/, media/ modules + @discordjs/voice dep
- Backend: cut voice/media/recordings oRPC routers, WS gateway-PCM auth + voice
  handlers, Redis bridge voice aggregation, redis-channels voice/media constants,
  config keys, moderation-types voice items, e2e voice/recordings suites
- Keep voice DB tables (destructive migration avoided); chatbot voiceRecordings
  tool + dashboard count remain as read-only historical data access

services/backend/package.json

@@ -20,7 +20,6 @@
     "typecheck": "tsc --noEmit"
   },
   "dependencies": {
-    "@discordjs/voice": "^0.19.2",
     "@orpc/server": "1.15.2",
     "axios": "^1.20.0",
     "dotenv": "^18.0.1",