Commit Graph
6 Commits
Author SHA1 Message Date
J3vbandClaude Fable 5 6afa9e974c refactor(server): thread context.Context through the db layer and all callers
Fixes all 109 golangci-lint findings (106 contextcheck, 1 gocritic,
2 gosec) that accumulated after D2 wired dbgen (whose queries take ctx)
under ctx-less db.DB wrappers while CI lint was quota-dead. No nolint
comments added; every finding fixed by genuinely threading context.

- db: all 138 hand-written db.DB methods take ctx first; the dbCtx()
  Background shim is deleted; raw Query/QueryRow/Exec/Begin use their
  Context variants; the four redundant ctx-less passthroughs removed.
  db.Auditor/WriteAudit gain ctx.
- Seams: permissions.Checker (DB iface, HasChannelPerm,
  RequireChannelAccess) and the service.Store interface mirror the new
  signatures (ws.EventStore and plugin.PluginStore already did).
- Callers: api/admin handlers use r.Context(); ws per-message paths use
  the connection ctx via DispatchV2; hub loops and startup wiring use
  context.Background(); service methods thread ctx where they have one
  and Background where no ctx exists. Public service surface reached by
  ctx-holding chains (PermissionService.HasChannelPerm/GetRoleForUser/
  RequireChannelAccess, message/dm/block/invite/profile methods) is now
  ctx-first.
- Detached (context.WithoutCancel) where cancellation would break an
  invariant, found by a 3-lens adversarial review of the diff:
  * voice-leave background retries (a dead webhook/connection ctx killed
    retry 2 before it ran, leaving ghost capacity-holding voice rows)
  * rollbackVoiceJoin's compensating delete (its trigger IS the cancel)
  * post-2FA-change DeleteOtherSessions and logout DeleteSession (the
    security tail of a committed change must not die with the request)
  * all api/ws audit writes (a banned user could suppress their own
    login_blocked_banned row by aborting the request mid-bcrypt)
  * admin backup VACUUM INTO (an interrupt left a truncated .db that
    the backup list presented as restorable)
  * post-commit message/edit refetches (a committed message must still
    fan out when the sender disconnects)
  * hub settings-cache refresh (one dead connection could pin stale
    values for the 30s TTL)
- gocritic rangeValCopy fixed (index iteration); gosec G306 excluded in
  config with justification (generated source must stay world-readable)
  instead of flipping genprotocol output to 0o600.

Verified: gofmt/vet, all four build-tag variants, full suite, deadlock
pass, full -race pass, golangci-lint 0 issues uncapped.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-23 17:03:52 +02:00
Claude 5f1d6fc287 refactor(server): remove the store abstraction layer (D3)
Deletes Server/store (SQLiteStore, MemStore, the composed Store
interface) and collapses to a single sqlc-backed db package, executing
the prior audit's P4 "single data layer" direction (finding #6).

SQLiteStore was a pure pass-through to *db.DB, so consumers now depend
on narrow interfaces that *db.DB satisfies directly:

  - service.Store   (service/datastore.go, renamed from store/store.go)
  - ws.EventStore   (ws/eventstore.go)
  - plugin.PluginStore (plugin/pluginstore.go)

The event- and plugin-KV methods that lived in the store's SQLite
implementation move into the db package (db/event_queries.go,
db/plugin_queries.go), keeping their raw-SQL form.

Tests: the MemStore-based unit tests now run against a real in-memory
SQLite db opened per-test with migrations applied, via package-local
seed helpers. Fault-injection tests embed a real *db.DB and override the
single method under test, preserving error-path coverage. Full server
suite and sqlc-verify are green.

Docs: audit finding #6 and A-2026-07-06 marked resolved; decisions D3
updated; architecture server.md / data-model.md diagrams and prose
updated to the api -> service -> db layering.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UA17KPvqGBX3XbXYnMf1rA
2026-07-19 16:33:58 +00:00
J3vb 9116a880a3 feat(phase-bc): pass 5 — pgdbgen, postgres EventStore/PluginStore, plugin hub wiring, OTel stack, reconnect DB tier
- Generate Server/db/dbgen/{events,plugins}.sql.go and full Server/db/pgdbgen/ (//go:build postgres gated)
- Implement PostgresStore EventStore and PluginStore methods in store/postgres.go
- Wire plugin host_events.go EventSink into hub broadcast path (SetPluginEventSink)
- Wire host_commands.go slash-command dispatcher: chat_command V1 handler + hub.SetPluginRegistry
- Add handlers_command.go + handlers_command_test.go for plugin slash-command dispatch
- Add reconnect_db_test.go: TestReconnect_BufferMiss_FallsBackToDBTier (cold-tier DB replay)
- Add otel-up/otel-down Makefile targets; docker-compose.otel.yml + prometheus.dev.yml
- Update PHASE_BC_LOCAL_TODO.md: mark in-session items complete; document remaining network-blocked steps
- Minor fixes: channel_handler access-control, router plugin handler wiring, service span instrumentation
2026-04-06 22:48:59 +02:00
J3vb 3d18ce4f99 fix: Go E2EE security hardening — key holder tracking, base64 loose validation, rate limits, test schema
- I-1: Add key holder election in Hub (lowest userID per channel); reject
  non-key-holder voice_e2ee_offer with NOT_KEY_HOLDER error
- I-2: Accept raw (unpadded) base64 in E2EE announce/offer handlers via
  decodeBase64Loose fallback
- I-6: Copy E2EE public key value while h.mu.RLock is held in getClientE2EEPubKey
- I-7: Lower loginRateLimitPerMinute from 60 to 5
- C-1: TOCTOU fix — target channel check held under same lock as client lookup
- C-2: Include is_key_holder bool in voice_token payload so client knows
  whether to initiate key distribution
- M-5/M-6: Add ErrCodeBadPayload/ErrCodeNotKeyHolder error constants
- Fix pre-existing api build errors: block_handler.go getUserFromContext,
  router.go RequirePermission arg count
- Add user_blocks table to all test DB schemas (ws, api DM)
- Add voice_e2ee_test.go and constants_test.go covering all fixes
2026-04-04 23:16:31 +02:00
J3vb 9e48e8d8e8 fix: security hardening — 11 findings across auth, WS, upload, admin, data
Security audit across all 11 sections (AUTH-001 through DATA-001) found
0 critical, 1 high, 7 medium, 15 low issues. This commit addresses:

- Add json:"-" to User.PasswordHash, User.TOTPSecret, Session.TokenHash
  to prevent accidental serialization of sensitive fields (M7)
- Add X-Content-Type-Options: nosniff to file serve responses (M5)
- Apply owner-only guard to backup list endpoint for consistency (M6)
- Persist rate-limit lockouts to SQLite so they survive restarts (M2)
- Normalize DM non-participant responses to 404 to prevent oracle (L3)
- Add explicit per-entry expiry check in partial auth Lookup/Consume (L1)
- Truncate unknown WS message type to 64 chars before echo (L6)
- Rate-limit ping handler to 2/sec per user (L7)
- Replace raw error strings in update handlers with generic messages (L15)
- Update 4 tests to match new 404 behavior for DM non-participant
2026-04-02 14:51:16 +02:00
jevbandclaude-flow c53d63da47 feat: comprehensive spec docs, test suite, E2E overhaul, and security hardening
Spec Documentation (18 files, 680KB):
- Expanded all 15 existing spec files with deep detail from source code
- Created 3 new specs: DM-SYSTEM, THEME-SYSTEM, RECONNECTION
- Created E2E-BEST-PRACTICES spec
- Audited all specs against source: fixed 50 errors

Unit Tests (143 new):
- Go: dm_queries_test (21), dm_handler_test (17), dm_handlers_test (18), ringbuffer_test (22)
- TS: dm-store (16), disposable (14), themes security (17), ws reconnection (8), dispatcher DM (2)

E2E Tests (22 mocked + 6 native specs):
- New: dm-system, theme-persistence, reconnection (mocked + native)
- Fixed 12 fake assertions, 18 hardcoded timeouts, 5 stale selectors
- Persistent fixture: login once per run instead of per test
- ensureLoggedIn with exponential backoff for rate limiting

Security Fixes:
- DM auth bypass: added IsDMParticipant to handleGetPins, handleSetPinned, handleSearch
- LiveKit InsecureVerifier replaced with PinnedVerifier (TOFU from shared cert store)
- IDOR leak: handleChatEdit/Delete now return opaque error codes
- CSS injection: added deny-list for dangerous CSS functions in themes
- BANNED error now triggers logout instead of infinite reconnect
- CredFree leak fixed: Windows credential memory freed before parsing
- Login lockout off-by-one: limit=9 so 10th failure triggers lockout

Stability Fixes:
- Rate limiter StartCleanup goroutine now started (prevents memory leak)
- Voice mute/deafen rate limiting added (2/sec, matching camera/screenshare)
- DM typing no longer echoes back to sender
- Accept loop spin protection (5 consecutive error limit)
- voice_config protocol drift resolved (3 missing fields added)
- Login rate limit set to 60/min (spec updated, 10-failure lockout is real protection)
- Hardcoded roleNameToId replaced with dynamic lookup from ready payload

Co-Authored-By: claude-flow <ruv@ruv.net>
2026-03-28 10:39:23 +01:00