Found while live-testing the realtime dictation:
- 'already active' lockout (real bug): the per-user slot was tied to the
connected socket lifetime and a stale/racing socket could leave the counter
stuck, so a fresh mic start was rejected. Now per-user single-session is
enforced purely by LATEST-WINS EVICTION — a new connect disconnects the user's
prior socket and frees its slot synchronously — and the user counter no longer
participates in the cap decision (it could only cause false lockouts). Also
free the slot when a start fails to open. The per-workspace cap is unchanged.
- #737: drop the separate sttRealtimeModel / sttRealtimeBaseUrl settings — realtime
dictation now reuses the existing STT model + base URL (the realtime WS endpoint
is derived from it server-side). Removed the fields from the DTO, types, settings
service, repo allowlist, and the settings UI. The STT 'Test endpoint' button is
now a single context-aware button (probes the realtime WS endpoint when realtime
is on, the batch endpoint otherwise), and the 'Request format' selector is
disabled while realtime is on (realtime always uses the OpenAI Realtime protocol).
- no-silent-loss: parse the OpenAI
conversation.item.input_audio_transcription.failed event (e.g. insufficient_quota,
bad model) and surface its concrete reason to the client instead of dropping it
silently — previously a per-item transcription failure produced 'no words' with
no explanation.
Tests: realtime suites green (gateway latest-wins eviction, parser .failed surfacing,
ai-settings reuse-STT-model); server + client tsc clean; workspace vitest 37 pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Implements all reviewer comments (code-review, red-team, and test-strategy
audit), accepting the recommended variants.
Server — realtime service (ai-realtime.service.ts):
- SSRF: pin the validated IP via a WebSocket `lookup` hook that re-checks every
resolved address with isIpAllowed (mirrors external-mcp buildPinnedDispatcher),
closing the TOCTOU/DNS-rebinding window; fix the misleading comment.
- no-silent-loss: on Stop, drain the in-flight segment (bounded 2.5s) and deliver
the final via onFinal before closing instead of dropping the tail.
- fail-closed deriveRealtimeUrl: a non-empty unparseable base now THROWS (no
silent api.openai.com fallback that would leak a self-hosted key); http://ws://
bases rejected (plaintext key). Path normalization preserved.
- parseUpstreamEvent keys the accumulator by item_id+content_index so GA segments
don't concatenate.
- inject a wsFactory seam for testing; also fix a latent bug — `import WebSocket
from 'ws'` resolved to undefined at runtime (no esModuleInterop) -> import=require.
- unref idle/max/drain timers.
Server — realtime gateway (ai-realtime.gateway.ts, session-limits.ts):
- reject revoked/disabled users and inactive sessions (mirror jwt.strategy:
findById+isUserDisabled + findActiveById) with NO counter increment.
- CSWSH: Origin allowlist (matching APP_URL, or no Origin for native clients)
before auth, no increment.
- extract SessionCounters (delete-at-zero, never negative) + pure canConnect
(both caps >= checked before any increment); document the per-process/in-memory
cap caveat (single-replica only).
Client:
- dictation-group: realtime final now inserts at the captured rangeRef SNAPSHOT
(not the live caret) and guards editor.isEditable; single-space separator.
- use-realtime-dictation/realtime-dictation-client: stop-during-acquisition tears
down the mic (no leak / button reset); reconnect re-emits start (double-start
guarded); interim ghost cleared on teardown; io() options de-duplicated.
- pcm16-worklet: flush the partial sub-frame tail on stop; one-pole anti-aliasing
low-pass before 48k->24k.
- extract shared mic-capture (acquireMicStream/mapGetUserMediaError, used by batch
+ realtime), pure DSP (pcm16-dsp.ts), and the session reducer/baseLanguageSubtag;
extract applyInterimMeta/clampRange/resolveUrl/appendFinalToDraft.
Tests + infra: +~150 server tests (deriveRealtimeUrl, parseUpstreamEvent branches,
openSession/lifecycle/timers/testConnection via fake ws, gateway auth/caps/no-leak,
realtime-test admin contract, AiSettings update/resolve, DTO boolean, SSRF deny)
and +~140 client tests (DSP property/edge, resampler continuity, framing, reducer,
mic-capture, RealtimeDictationClient/MicButton, ProseMirror interim regression +
history guards, appendFinalToDraft, resolveKeyField, route contract). Added
@vitest/coverage-v8. CHANGELOG [Unreleased] entry incl. the single-replica caveat.
Review: APPROVE WITH SUGGESTIONS (no critical/regression); applied the drain-timer
unref. Server tsc clean + 358 tests; client tsc clean + 201 tests; vite build ok.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Layer an optional realtime speech-to-text path on top of the existing
batch dictation, so transcribed text appears as the user speaks.
Transport A2: browser <-> our server (Socket.IO `/ai-realtime`) <->
OpenAI Realtime (raw ws). The provider API key never leaves the server;
the upstream URL is SSRF-checked before connecting; the gateway enforces
the dictation+dictationRealtime gate, cookie-JWT auth and per-user/
per-workspace concurrency caps. Implemented against the GA (2026) OpenAI
Realtime transcription contract (session.update / audio.input.format /
server_vad), not the now-removed beta shape.
Editor UI B2: interim text is shown as a meta-only ProseMirror ghost
decoration (no Yjs/history noise); only completed segments are committed.
Chat shows interim as a dimmed tail. The mic button switches realtime vs
batch by the workspace flag; batch remains the default and fallback.
Server:
- AiRealtimeService (upstream ws proxy, normalized events, idle/max-
duration timeouts, idempotent teardown) + parseUpstreamEvent unit tests
- AiRealtimeGateway (Socket.IO `/ai-realtime`) wired into AiChatModule
- admin-gated POST /ai-chat/realtime/test connectivity probe
- config: settings.ai.dictationRealtime + provider sttRealtimeModel/
sttRealtimeBaseUrl (realtime key reuses sttApiKey; no new secret)
Client:
- pcm16 AudioWorklet (24kHz mono PCM16), RealtimeDictationClient,
use-realtime-dictation hook (status/start/stop/cancel + onInterim/onFinal)
- RealtimeMicButton + dictation-interim ProseMirror decoration
- editor/chat integration + AI settings UI (toggle, model, test endpoint)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Pre-merge review follow-up for the parseNodeArg dedupe (PR #114):
- Restore docs/backlog/ai-chat-tool-definitions-duplicated.md instead of
deleting it: it still tracks open debt (unified spec registry + ProseMirror
<-> Markdown converter unification) that this branch defers, and
docs/git-sync-plan.md links to its converter section. Mark the node-arg
quirk as done and add a Progress section.
- Reword the in-app helper header from "byte-for-byte" to "behaviorally
identical": the two copies differ in comments/quote style; only the logic,
throw messages and branch order match.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The fail-closed limiter behavior (#62 primary item) already shipped; this
finishes the issue by lowering the default hourly per-workspace cap from 300
to 100 to better fit real anonymous-assistant load. Still overridable via
SHARE_AI_WORKSPACE_MAX_PER_HOUR.
- public-share-workspace-limiter.ts: SHARE_AI_WORKSPACE_MAX_PER_WINDOW 300 -> 100.
- .env.example: documented default + example value 300 -> 100.
- public-share-chat.spec.ts: update the default-cap assertion to 100.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Implements Option 2 of #93. The restricted branch of broadcastPageMoved
previously resolved its audience twice — emitToAuthorizedUsers and
emitDeleteToUnauthorized each ran an independent fetchSockets +
getUserIdsWithPageAccess — leaving a race window between the two snapshots
where a socket could receive both the move and the delete (leak) or neither
(lost compensating delete).
- ws.service.ts: add emitMoveWithRestrictionSplit() that takes ONE socket
snapshot and ONE authorization resolution, then partitions the room:
authorized users get the moveTreeNode, everyone else (unauthorized +
anonymous) get the compensating deleteTreeNode. Disjoint + complete by
construction. Remove the now-unused emitToAuthorizedUsers /
emitDeleteToUnauthorized; keep private broadcastToAuthorizedUsers (still
used by emitRestrictedAwareToSpace).
- ws-tree.service.ts: broadcastPageMoved restricted branch now drives move +
delete from the single method.
- specs: assert the single method is used and that fetchSockets /
getUserIdsWithPageAccess are each called exactly once (single snapshot);
re-route ws-service.spec to emitTreeEvent after the method removal.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Builds the deferred integration tests from docs/backlog/feature-test-coverage-
deferred.md that needed real infra (a test Postgres + real Redis) which the repo
lacked. Runs against an isolated, auto-created docmost_test database and Redis
logical DB 15 — never the dev data.
Harness (apps/server/test/integration/, run via new `pnpm --filter server test:int`
=> jest --config test/jest-integration.json; default unit `jest` is untouched and
excludes these via the *.int-spec.ts name + rootDir):
- db.ts: buildTestDb() mirrors database.module.ts exactly (PostgresJSDialect,
CamelCasePlugin, bigint to:20/from:[20,1700] parsing) + minimal seed helpers.
- global-setup.ts: DROP/CREATE docmost_test, CREATE EXTENSION vector, migrate to
latest via Kysely Migrator (fails loud on any errored migration).
- global-teardown.ts: closes the pool.
Coverage (5 suites, 16 tests, all green against live PG+Redis):
- WorkspaceRepo.updateSetting: jsonb-merge persists htmlEmbed without clobbering
sibling ai/sharing namespaces (the kill-switch write half).
- AiAgentRoleRepo: soft-delete exclusion, cross-workspace tenant isolation,
duplicate (name,workspace) -> 23505, name reusable after softDelete (partial
unique index WHERE deleted_at IS NULL), same name across workspaces allowed.
- page_template_references: deleting either source or referenced page cascades
the link row (onDelete cascade) — real FK, not mocked.
- PublicShareWorkspaceLimiter vs REAL Redis: real ioredis EVAL of the sliding-
window Lua — max boundary (3 admit / 4th deny), re-admit after the window
slides, same-ms distinct members. Catches Lua bugs a FakeRedis cannot.
- AiChatRepo.findByCreator: role-badge join (enabled->badge; soft-deleted or
disabled role -> null).
Review: APPROVE; applied its two hardening suggestions (fail loud on errored
migration result even without a top-level error; TEST_REDIS_URL override + ping
preflight). tsc clean; unit run excludes int-spec (verified).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
First, safe step of docs/backlog/ai-chat-tool-definitions-duplicated.md: the
"node may be a JSON object OR a JSON string" quirk was hand-copied at 6 tool
sites. Extract it into a single parseNodeArg() helper per package and call it at
every site. Behavior-preserving — each site's throw message is byte-identical
(patch/insert: 'node was a string but not valid JSON'; update_page_json: 'content
was a string but not valid JSON'); no tool name/description/schema changed.
Two helper copies (packages/mcp/src/lib/parse-node-arg.ts and
apps/server/src/core/ai-chat/tools/parse-node-arg.ts) are intentional: the
ESM-only @docmost/mcp cannot be imported by the CommonJS server (it is loaded at
runtime via the Function('import()') trick), so runtime code cannot cross that
boundary by a normal import. Each copy is now the single source within its
package (6 inline copies -> 2 helpers). packages/mcp/build rebuilt in sync.
Tests: parse-node-arg.spec.ts (server, Jest) + parse-node-arg.test.mjs (mcp,
node:test) — object passthrough, valid-string parse, invalid-string throw with
the right message. Server tsc clean; mcp suite 254 pass; agent structural-edit
path verified live in-browser (agent inserted a node, persisted to the doc).
Deferred (documented for the record, since the backlog doc is removed with this
commit): the FULL transport-agnostic tool-spec registry (one name+schema+
description per tool shared by both transports) and deriving DocmostClientLike
from the real client type. Both are blocked by the current architecture, not by
effort: (1) @docmost/mcp ships no type declarations and is ESM-only, so a
type-only derivation needs declaration emission + tsconfig path wiring, and the
real client's precise return types break the in-app tool test stubs (attempted,
reverted to keep tsc green); (2) the two transports intentionally DIVERGE in tool
NAMES (snake_case x38 vs camelCase x41), membership (in-app adds getCurrentPage/
listSidebarPages, omits delete_comment/image tools) and model-facing
DESCRIPTIONS, so a unified registry would change behavior on BOTH the agent and
external MCP clients and needs its own verification pass. This is forward-looking
debt (the code is correct today), to be done incrementally.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The 'current page' feature (client useMatch openPage + server getCurrentPage
tool + system-prompt injection) was already implemented & merged; this backfills
its missing test coverage and removes the completed backlog doc.
- extract pure resolveCurrentPageResult(openedPage) into current-page.util.ts
(byte-identical to the prior inline getCurrentPage tool body) so it is
unit-testable without the dynamically-imported ESM Docmost client; the tool
now delegates to it.
- current-page.util.spec.ts: 7 cases (null/undefined/no-id/empty-id/full/no-title).
- ai-chat.prompt.spec.ts: +8 cases for the openedPage context line (title+pageId
present, Untitled fallback for blank/whitespace title, no line when absent/blank
id, and sandwich ordering before the trailing safety block).
Verified live in-browser: client sends openPage{id,title} on a page and null
off-page; the agent invokes getCurrentPage and answers with the real title+id.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
DTO: trackerHead @IsString/@MaxLength(20000) + htmlEmbed @IsBoolean accept/reject
cases. CASL: a non-admin updating trackerHead/htmlEmbed gets ForbiddenException
(update not called); owner/admin proceed. Audit: a no-op trackerHead re-save
doesn't enter the audit diff.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Extract the admin trackerHead <head> injection into a pure injectTrackerHead()
and test it: a snippet containing $&/$$/backtick-dollar survives BYTE-FOR-BYTE
(pins the function-replacer fix), empty/whitespace/undefined and a missing </head>
leave the html unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Relocate resolveTrustProxy from main.ts (untestable — bootstraps on import) to
integrations/environment/trust-proxy.util.ts and import it back. Unit-test every
branch (empty/undefined -> safe loopback/private default; true/false; hop count;
trim; CIDR/negative passthrough) so a regression can't silently re-open the XFF
spoofing hole (#61).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Covers the #63 guard: a message with a non-text part -> 400 'Unsupported message
content'; a message mixing text + a non-text part still 400s (before the 413
size check).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds the missing tests for the #67 guard: self-move and a destination inside the
moved page's subtree both throw BadRequestException before updatePage; a
legitimate move proceeds. Mocks pageRepo + spies getPageBreadCrumbs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- ai-chat: drop the unused pagePermissionRepo injection from
PublicShareChatToolsService (its only use moved into
ShareService.resolveReadableSharePage); update all 5 positional
test construction sites to match the 3-arg constructor.
- env: correct the anonymous share-AI per-workspace cap comment —
the limiter FAILS CLOSED on Redis failure (#62), not open.
- docs: sync README.ru.md with README.md — move "Page templates"
from Planned to Done and drop the dead plan-doc link.
Remaining test-coverage gaps tracked as #102, #103, #104, #105, #106.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Reconcile the diverged develop (13 ahead / 20 behind) with gitea/develop.
Conflict resolution — html-embed: keep the local sandboxed-iframe model
(opaque-origin srcdoc, no role-gating) and supersede gitea's same-origin
strip/kill-switch hardening (#26/#28/#29/#30). The 4 conflicted html-embed
source files resolve to the local version; the 3 strip-era spec files stay
deleted. The strip apparatus (stripDisallowedHtmlEmbedNodes,
collectHtmlEmbedSources, canAuthorHtmlEmbed, htmlEmbedAllowed) is fully gone.
Integrate gitea's page-templates / page-embed work (#31-#40) cleanly.
Fix an auto-merge arity mismatch: two new gitea page-template specs
constructed TransclusionService with the pre-sandbox 11-arg signature; drop
the trailing workspaceRepo argument to match the reduced 10-arg constructor.
Verified: server + client tsc --noEmit clean; jest (html-embed + transclusion)
14 suites / 119 tests passing.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Deduplicate the "save a workspace setting" plumbing shared by HtmlEmbedSettings
and TrackerSettings (workspace atom read, isLoading state, updateWorkspace + atom
merge forcing settings[key], success/error notifications) into a new
feature-scoped hook useWorkspaceSetting(key).
- Each component keeps its own interaction model: html-embed is an optimistic
toggle with revert-on-failure; tracker is edit-then-save on an explicit button.
- Unify error handling on the better pattern: surface err.response?.data?.message
and use console.error (html-embed previously used console.log + a generic message).
No user-facing behavior change; client typecheck clean.
Test-coverage follow-ups (untested trackerHead injection in ShareSeoController and
the no-op audit branch) tracked in #100.
isBlocked was checked synchronously but recordFailure ran only AFTER the bcrypt
awaits, so N concurrent /mcp Basic requests for one email all slipped past the
threshold. Add FailedLoginLimiter.tryReserve (atomic synchronous check+increment)
+ release (undo), and reserve all 3 keys BEFORE any await so the (threshold+1)-th
concurrent attempt is rejected before its bcrypt runs. The reservation IS the
failure record (post-await recordFailure removed -> counted exactly once). Non-
credential early throws (missing workspace, SSO/MFA gate) and business errors
release the reservation so they don't burn a victim's budget; success clears.
Tests prove login() runs exactly threshold times under concurrency and that
gate/config rejects don't consume budget.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The three collect*FromPmJson collectors shared the same recursion (and the #55
depth cap) but were copy-pasted, so a future edit could diverge them. Extract a
generic collectNodes(doc, {type, map, key, lastWins, skipChildrenOfType}) and
reimplement all three on it, byte-output-identical (transclusions last-wins;
references/embeds first-wins + transclusionSource skip). Documents (not removes)
the write-only page_template_references graph and the near-duplicate client
lookup-context as tracked follow-ups, per the issue's guidance.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The '(shareId,pageId) -> usable non-restricted page in THIS share' boundary was
written as 3 must-be-identical async sequences. They weren't: the chat funnel
omitted an explicit page.deletedAt check (latently safe via getShareForPage's
CTE) and layered isSharingAllowed separately. Add ShareService.resolveReadable-
SharePage(shareId,pageId,workspaceId) running the single canonical sequence
(getShareForPage -> id match (skipped when null) -> findById -> !deletedAt ->
!hasRestrictedAncestor) returning {share,page}|null; getSharedPage, the funnel,
and the getSharePage tool all use it. hasRestrictedAncestor now lives in the one
method no caller can skip; the funnel still returns uniform 404s and keeps
isSharingAllowed. Adds a direct security-invariant test.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
resolveRoleForRequest and resolveShareRole duplicated the security invariant
'role exists, not soft-deleted, enabled, workspace-scoped, else null'. Move it to
AiAgentRoleRepo.findLiveEnabled(id, workspaceId) (deletedAt IS NULL + enabled +
workspace scope) and have both services call it, preserving each one's roleId
derivation + null handling. (describeProviderError half of #95 was done earlier.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
emitTreeEvent and emitCommentEvent were byte-identical (same room resolution,
spaceHasRestrictions gate, hasRestrictedAncestor, authorized-only vs broadcast
fallback). Collapse the body into one private emitRestrictedAwareToSpace; both
stay thin wrappers with unchanged signatures, so the restriction-routing gate
lives in exactly one place. Adds coverage for the comment entry point.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
McpService.enforceBasicLoginGate re-implements AuthController.login's pre-token
SSO/MFA gate; silent drift would re-open the bypass. Add an AST contract test
(comments stripped) asserting BOTH method bodies contain validateSsoEnforcement,
the EE-MFA require, and checkMfaRequirements — so dropping the gate from either
side fails CI. Test-only (no core/auth refactor).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The 4-step html-embed gate (feature-enabled AND role-allowed -> stripHtmlEmbedNodes)
was replicated across call-sites, pinned only by brittle source-regex tests. Add
stripHtmlEmbedIfNotAllowed(json, {featureEnabled, role, onStrip}) and migrate the
5 plain strip-all sites (collab handler, page create+duplicate, both import paths,
transclusion) to it, each keeping its own feature/role resolve + log via onStrip.
Left the 2 sites with different semantics: persistence.extension (#29 preserve-
admin) and share.service (feature-only kill-switch, no role gate). Real unit tests
replace the regex pins; behavior identical.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
16 suites were disabled via testPathIgnorePatterns due to two root causes: lib0
ESM not transformed (the @hocuspocus/server -> lib0/decoding.js chain) and stock
'should be defined' specs built via Test.createTestingModule without providers.
Add lib0 to transformIgnorePatterns; convert the 14 DI placeholders to direct
new X(...) instantiation with stub deps (keeping a real construct smoke test);
re-enable the suites. Also updates the public-share limiter test to assert the
fail-closed behavior from #62 (surfaced now that the suite runs). Full server
suite: 67 passed, 689 tests.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The shared MCP_TOKEN guard moved from 'Authorization: Bearer <MCP_TOKEN>' to the
X-MCP-Token header (Authorization is now per-user Basic/Bearer), silently breaking
existing /mcp clients. Document it as a Breaking Change in CHANGELOG (reconfigure
to X-MCP-Token). Add a once-per-process migration warning: when MCP_TOKEN is set,
no x-mcp-token is present, and Authorization carries the old 'Bearer <MCP_TOKEN>',
log a hint to migrate — without changing the auth decision (still rejected) or
logging the token value.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Assessment of the page-embed depth/cycle cap: the server /pages/template/lookup
returns FLAT single-level content and does NOT recurse into embedded pages — the
recursive expansion + the PAGE_EMBED_MAX_DEPTH cap are entirely client render
concerns, and a scripted client is already bounded by the per-user throttle
(30/min) + the ArrayMaxSize(50) per-call cap. So no server-side depth guard is
needed; documented at lookupTemplate so future readers don't add a redundant one
or assume server recursion exists.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
handleMessage became a no-op and PageWsListener intentionally ignored
PAGE_UPDATED, so a rename/icon change (client operation:updateOne) was no longer
rebroadcast -> other clients saw stale title/icon in the sidebar+breadcrumbs
until a reload (create/duplicate/restore were covered; updateOne regressed).
Add a server-authoritative onPageUpdated handler: PageService.update detects a
real title/icon change (DTO carries the field AND value differs; no-op/content-
only saves excluded) and attaches a treeUpdate snapshot to PAGE_UPDATED; the
listener broadcasts a tree updateOne via the restriction-aware emitTreeEvent
(so a restricted page's title never leaks). Content-only saves attach nothing.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
sidebar-pages-tree.spec tested a LOCAL COPY of the tree-shaping (so a regression
in the real getSidebarPagesTree was invisible) and justified it with a false
jest-config claim (the ^src mapping exists). Extract the pure shaping into
shapeSidebarPagesTree(); the service now calls it and the spec imports the REAL
helper. Behavior unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The concurrent-soft-delete guard was already covered; add the missing assertion
that update() returns toView(updated) from the post-update re-fetch (full
AgentRoleView shape, distinct second findById row), so a regression returning the
stale pre-update view or leaking columns is caught.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The anonymous share page-fetch tool's positive branch (sanitize via
updatePublicAttachments then jsonToMarkdown before returning to the model) was
untested, so a dropped/reordered sanitizer would ship a comment-mark/raw-
attachment leak with green tests. Add a positive-branch test pinning the
sanitizer call + that markdown derives from sanitized content, and a soft-deleted
test asserting a generic error with no content fetch.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
invalidateSpaceRestrictionCache has no callers because no restriction-mutation
path exists yet (PagePermissionRepo mutators are uncalled; there is no
restrict/grant/revoke endpoint), so the 30s spaceHasRestrictions cache could
serve a stale 'no restrictions' verdict. Until a mutation endpoint exists to
wire the direct invalidation, lower the TTL (30s -> 3s) to bound the worst-case
window; the invalidation primitive is kept for that future endpoint.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
chatModel was a free string accepted with empty/garbage values, failing only at
runtime as a provider 503; tighten it (trim + non-empty + max 200). Driver was
already @IsIn(AI_DRIVERS). Collapse the client driver list to one AI_DRIVER_VALUES
source and add a contract test that reads the server AI_DRIVERS and fails on
client/server drift.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Follow-up fixes to the htmlEmbed-sandbox / trackerHead change:
- share-seo: inject trackerHead via a function replacer so `$`-sequences
($&, $', $`, $$) in the admin snippet are inserted literally instead of
being treated as String.replace substitution patterns; warn when the
</head> marker is absent instead of silently skipping injection.
- mcp: register a passthrough `htmlEmbed` node in the schema mirror so an
AI/MCP edit of a page containing an embed no longer throws
"Unknown node type: htmlEmbed" in TiptapTransformer.toYdoc.
- editor-ext + client: treat a non-finite `data-height` as auto (null) so a
crafted/corrupted height cannot disable auto-resize or yield a NaN iframe
height; extract a shared clampHeight helper.
- client: rename render-raw-html.{ts,test.ts} -> html-embed-sandbox.{...} and
shouldExecute -> shouldRender so the seam name matches the sandbox model.
- client: i18n the iframe title; surface the real error reason in
tracker-settings (console.error + err.response.data.message).
- docs: note hasHtmlEmbedNode is now a test-only helper; add an Unreleased
CHANGELOG entry; drop the dangling "arbitrary HTML embed" planning-doc ref.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
collectPageEmbedsFromPmJson (and the sibling collectors/remap) recursed with no
guard, so a pathological/cyclic non-JSON input could stack-overflow (RangeError).
Add a depth cap (1000, far above any real doc nesting) so such input is handled
gracefully. Normal documents are unaffected. Updates a stale test that asserted
the old throwing behavior.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
trustProxy was unconditionally true, so req.ip came from a client-forgeable
X-Forwarded-For and the per-IP throttles (share-AI, /mcp brute-force) were
spoofable. Make it env-configurable (TRUST_PROXY) with a safe default that
trusts XFF only from loopback/private proxies, documented in .env.example.
NOTE: this changes the default from trust-all; deployments whose proxy is on a
public IP must set TRUST_PROXY (caveat documented).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A custom AI-role's text preceded the only SAFETY_FRAMEWORK block and replaced
the persona, so a jailbreak in the role text sat before the safety rules.
buildSystemPrompt now emits SAFETY both before AND after the persona, with the
role/persona delimited as lower-trust (<role_persona note=...>); the default
persona is sandwiched too. Context (currently-viewing-page) preserved.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
#67: movePage didn't check the destination wasn't the page itself or inside its
own subtree, so MCP/REST/agent/fast-drag could persist+broadcast a cycle. Reject
before the update (self-parent, or moved page among the destination parent's
ancestors via getPageBreadCrumbs).
#64: movePage emitted PAGE_MOVED from a stale pre-read even when the row didn't
change / was concurrently deleted (phantom move). Gate the emit on
updateResult.numUpdatedRows !== 0n. Both are movePage hardening in one method.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
#60: streamText had no maxOutputTokens, so one anonymous request could run up
the provider bill. Add maxOutputTokens (env SHARE_AI_MAX_OUTPUT_TOKENS, default
512) via resolveShareAiMaxOutputTokens().
#95: the anonymous path hand-built error strings, diverging from the unified
describeProviderError format used on the authenticated path; both onError blocks
now call describeProviderError so a share reader sees 402/429/503 causes in the
same form (and the stack is still logged). Both changes are in this one file and
share hunks, hence one commit.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A user with password=NULL passed the missing/disabled guard and reached
comparePasswordHash(pw, null), which native bcrypt rejects -> 500 on
/api/auth/login and, on /mcp, a leaky 401 that the brute-force limiter ignored
(enumeration oracle + limiter evasion). Treat a null/empty password like a
missing user in verifyUserCredentials (dummy compare for timing parity + unified
CREDENTIALS_MISMATCH_MESSAGE) and reject early in changePassword before bcrypt.
Contract spec asserts the null-password guard.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
MAX_SHARE_MESSAGE_CHARS only counted text parts, so a forged non-text part
(tool-result/file/data) bypassed the cap and bloated the model input
(token-DoS); convertToModelMessages would also expand a forged tool-result. The
anonymous path runs no tools, so a client non-text part is never legitimate —
reject any message with a non-text part (isTextUIPart) before the size check.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The per-workspace anonymous share-AI cost cap failed OPEN on a Redis error
(return true => admit), so a Redis outage removed the cap entirely (unmetered
billable anonymous calls). The feature is optional, so unavailability is
harmless: fail CLOSED (return false => controller 429s) instead.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Convert the htmlEmbed node from same-origin raw-HTML execution to a sandboxed
iframe (sandbox="allow-scripts allow-popups allow-forms", no allow-same-origin,
srcdoc) with postMessage auto-resize (validated by event.source) and an optional
manual height attr. The block now runs in an opaque origin and cannot reach the
viewer's cookies/session/API, so it is safe for any member.
Because the block is now harmless, remove the entire admin/role gating apparatus:
drop htmlEmbedAllowed/canAuthorHtmlEmbed/stripDisallowedHtmlEmbedNodes/
collectHtmlEmbedSources and every role-based strip on the write paths (collab
REST/MCP + socket, page create/duplicate, import x2, transclusion unsync), along
with the now-unused WorkspaceRepo/UserRepo injections and the PageService.create
callerRole param. Keep one strip: prepareContentForShare still removes htmlEmbed
on the anonymous public-share read path when the workspace master toggle is OFF.
The workspace settings.htmlEmbed toggle is now a plain feature switch (gates the
slash-menu and share rendering); when ON the block is available to all members.
Add settings.trackerHead: an admin-only raw HTML/JS analytics snippet injected
verbatim into the <head> of public share pages only (ShareSeoController), for
trackers that genuinely need same-origin. Admin-gated via the existing CASL
Manage/Settings ability; never injected into the authenticated app shell.
Closes security-review findings #1, #2, #4, #5, #10 (and #3 as a security issue).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Harden the anonymous public-share AI assistant against token-cost abuse
before exposing it to the internet:
- Add an env-tunable per-request output ceiling (maxOutputTokens) to the
public-share streamText call so one anonymous request cannot run up the
provider bill even if the per-IP throttle is evaded. New
resolveShareAiMaxOutputTokens() / SHARE_AI_MAX_OUTPUT_TOKENS_DEFAULT
(env SHARE_AI_MAX_OUTPUT_TOKENS, default 512), mirroring
resolveShareAiWorkspaceMax().
- Flip the per-workspace cost limiter to FAIL CLOSED on Redis failure
(was fail-open): if Redis is unavailable we cannot prove the workspace is
under its cap, so deny rather than admit an unmetered, billable call.
- Update the limiter spec (fail-open -> fail-closed) and add resolver tests;
document both knobs in .env.example.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
