- F1: document AI_CHAT_DEFERRED_TOOLS in .env.example (AI_* section) — default
ON = deferred loading (compact catalog + loadTools), =false restores the old
"all tools always active" behavior.
- F2: integration test of the ON path in ai-chat-stream.int-spec.ts — a deferred
tool activated via loadTools is active on the SAME turn's next step but a fresh
turn starts cold (CORE + loadTools only), proving the per-turn activatedTools
Set does not leak across turns/chats. Drives the real streamText loop with a
MockLanguageModelV3 and inspects recorded per-step activeTools-filtered tools.
- F3: replace the magic toHaveLength(28) in tool-tiers.spec.ts with a two-way
partition against the LIVE in-app toolset (AiChatToolsService.forUser keys):
every non-core tool must appear in buildInAppDeferredCatalog and every catalog
entry must map to a real non-core tool — so a future tool forgotten in
INLINE_TOOL_TIERS fails the suite instead of silently vanishing from the agent.
No production logic change (mechanism was already reviewed correct).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Per maintainer directive (#119 comment): land the canonical converter on the
git-sync branch so sync is tested on the real format, NOT a dead legacy copy.
#119 itself stays FROZEN (не вливается) — this only merges develop in.
Resolutions (all git-sync converter conflicts → develop; engine kept as-is):
- Dropped the branch's legacy `packages/git-sync/src/lib/*` converter — the
converter now lives solely in `@docmost/prosemirror-markdown` (#293); the
engine (pull/push/stabilize/index) only switches its imports to the package
(no logic change, verified by diff).
- Removed the branch's orphaned converter tests + fixtures under
`packages/git-sync/test/` (their coverage moved to the package's own test
suite on develop); git-sync/test now holds engine tests only.
- .gitignore / Dockerfile / test.yml / AGENTS.md: unioned — build/ ignored for
every package; Dockerfile COPYs both prosemirror-markdown/build (mcp+git-sync
runtime) and git-sync/build (git-sync's runtime consumer lands on this branch);
CI builds prosemirror-markdown before git-sync/mcp.
- pnpm-lock.yaml regenerated for the merged workspace.
Branch adaptations to canon (server-side tests only — converter untouched, per
the guardrail that converter fixes go to the package on develop, fixtures-first):
- git-sync-converter-gate.spec.ts: heading textAlign and image width/height now
round-trip via the canon trailing-comment forms (#9 `<!--attrs {...}-->`, #4
`<!--img {...}-->`) instead of the old HTML-tag forms — expectations flipped to
the real canon output. RESIDUAL: canon #4 does not yet carry image `align`
(documented as a known divergence; fix belongs in the package on develop).
- schema-attribute-contract.spec.ts: the schema mirror moved from
`@docmost/git-sync/lib/docmost-schema` to `@docmost/prosemirror-markdown`;
import + jest source-mapper updated.
Verified: prosemirror-markdown/git-sync/mcp build clean; git-sync corpus green;
server `tsc --noEmit` 0; gate + schema-attribute-contract specs 32/32.
The in-app AI agent shipped all ~41 tool schemas on every model step. This
adds a two-tier catalog: core tools (frequent or one-line) stay always-active;
the rest are advertised as a compact catalog and their full schema is fetched
on demand via the loadTools meta-tool, wired through ai@6 prepareStep's
per-step activeTools.
- tools/tool-tiers.ts: CORE_TOOL_KEYS, INLINE_TOOL_TIERS, applyLoadTools,
catalog builders (+ tool-tiers.spec.ts, 13 cases).
- ai-chat.service.ts prepareAgentStep: returns activeTools =
[...CORE_TOOL_KEYS, loadTools, ...activatedTools]; per-turn activated Set.
- ai-chat.prompt.ts: buildToolCatalogBlock renders the deferred catalog.
- mcp/tool-specs.ts: tier + catalogLine metadata (external snake_case /mcp
transport unchanged).
- EnvironmentService.isAiChatDeferredToolsEnabled(): AI_CHAT_DEFERRED_TOOLS,
default ON per issue intent (kill-switch =false restores old behavior).
Gate: server ai-chat 631/631, tool-tiers 13/13, mcp 472/472, tsc clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F4 [critical] — the anti-join `DELETE … WHERE NOT EXISTS(child)` was still racy
under Postgres READ COMMITTED: a reply INSERT holds FOR KEY SHARE on the parent;
the DELETE's start snapshot doesn't see the uncommitted child (NOT EXISTS true),
blocks on the reply's lock, and when the reply commits the parent was only LOCKED
(not modified) so EvalPlanQual does NOT re-check → the DELETE proceeds and CASCADE
destroys the just-committed reply. Replaced with a transaction: SELECT the parent
FOR UPDATE (conflicts with the reply's FOR KEY SHARE → serializes the concurrent
reply), re-check for a child with a FRESH statement in the same tx (a new RC
snapshot sees a just-committed reply), delete only if still childless (return 1)
else return 0 (caller resolves). The FOR UPDATE lock is held to end-of-tx so no
reply can insert between the re-check and the delete. Signature unchanged, so the
service + its mocked unit tests are untouched; docstrings updated.
F5 [warning] — the client Dismiss button was gated only on canComment, but the
server now gates dismiss on owner-or-space-admin, so a non-owner non-admin saw a
button the server 403s. `canShowDismiss` now also requires
`isOwnerOrAdmin = currentUser?.user?.id === comment.creatorId || userSpaceRole ===
"admin"` (the same gate the comment delete-menu already uses); threaded into both
call sites.
F6 [warning] — added a REAL-DB int-spec
(apps/server/test/integration/comment-delete-if-childless.int-spec.ts, + a
createComment seeder): (a) childless → returns 1, row gone; (b) committed reply →
returns 0, parent+reply survive; (c) CONCURRENCY — a second connection inserts a
reply (FOR KEY SHARE) and commits mid-operation while deleteCommentIfChildless
blocks on FOR UPDATE → asserts it returns 0 and both rows survive (a blind
anti-join would lose the reply here). Ran against live Postgres — 3/3 pass.
server tsc clean; comment jest 53 + int-spec 3 (live Postgres) pass. client tsc
clean; comment vitest 56 pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Maintainer escalation decision (B) + reviewer findings on the ephemeral-
suggestion PR.
Authz (decision B): POST /comments/dismiss-suggestion now gates the destructive
branch on owner-OR-space-admin, mirroring POST /comments/delete exactly (same
SpaceCaslAction.Manage / SpaceCaslSubject.Settings, same owner short-circuit,
same ForbiddenException). A non-owner non-admin who tries to dismiss another's
childless suggestion gets Forbidden before the service runs. Apply stays on
canEdit (accepting an edit is the editor's semantics), unchanged.
F1 [blocking] — atomic conditional delete closes the hasChildren→delete race.
New repo `deleteCommentIfChildless(id)` runs a single
`DELETE FROM comments WHERE id=:id AND NOT EXISTS (SELECT 1 FROM comments child
WHERE child.parent_comment_id = comments.id)` (verified by compiling the Kysely
expression to SQL — the correlated subquery references the OUTER comments.id).
deleteEphemeralSuggestion strips the mark first, then the conditional delete: if
it removed the row → commentDeleted + outcome 'deleted'; if a reply raced in
(0 rows) → fall back to resolveComment (outcome 'resolved') so the discussion and
the new reply survive. No reply can be cascade-deleted anymore.
F2 [warning] — the apply/dismiss onError success-noop is narrowed from 404||400
to 404 ONLY. A 400 means the comment is ALIVE (apply's 400 = the thread was
resolved-not-applied), so it now shows a real error (surfacing the server
message) and KEEPS the comment in cache instead of a false "applied" + dropping a
live thread.
F3 [suggestion] — the 404-race client tests assert the success toast fired.
Tests: server — dismiss authz (owner ok / non-owner-non-admin Forbidden /
space-admin ok), the delete→resolve race (hasChildren=false but conditional
delete returns 0 → resolve, no commentDeleted), delete-path asserts switched to
deleteCommentIfChildless; client — apply-400 and dismiss-400 (kept in cache, red,
not success) + the toast assertions.
server tsc clean, comment+collaboration jest green; client tsc clean, comment
vitest 54 passed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Agent suggestion-edits (comments with suggestedText, #315) piled up: Apply
auto-resolved the thread, cluttering the resolved tab, and the anchors stayed in
the document. Make them ephemeral: resolving (Apply OR the new Dismiss) makes the
comment DISAPPEAR — hard-delete + remove the Yjs `comment` mark — UNLESS the
thread has replies, in which case resolve it (preserve the discussion). Manual
Resolve is unchanged. Scope: only comments with `suggestedText`.
Server:
- New collab event `deleteCommentMark` (collaboration.handler) mirroring
resolveCommentMark, wiring the existing removeYjsMarkByAttribute to strip the
anchor from the doc.
- `finalizeAppliedSuggestion` forks on `hasChildren`: replies → apply + resolve
(outcome 'resolved'); none → apply + hard-delete + mark removal (outcome
'deleted').
- New `dismissSuggestion` (validates top-level + suggestedText + not applied/not
resolved) with the same fork; permission `canComment` (NOT canEdit — dismiss
doesn't change page text); audit COMMENT_SUGGESTION_DISMISSED. New
POST /comments/dismiss-suggestion; apply stays canEdit.
- Both return `{ outcome: 'deleted' | 'resolved' }` so the client picks the
optimistic action.
Data-integrity (review F1): the shared `deleteEphemeralSuggestion` removes the
anchor mark FIRST and FATALLY, then deletes the DB row only on success. The row
delete is irreversible, so a mark-removal failure — including the
COLLAB_DISABLE_REDIS "no live instance" hard-error — must abort the whole
operation (→ 5xx, repeatable) rather than swallow the error and leave a permanent
orphan anchor pointing at a deleted comment. `deleteCommentMark` is no longer
best-effort (unlike resolve, where the row is kept and a failed mark is
recoverable).
Client:
- `canShowDismiss` (canComment) alongside `canShowApply` (canEdit); a "Dismiss"
button next to Apply in the suggestion block.
- `useApplySuggestionMutation`/`useDismissSuggestionMutation` reconcile the cache
on `outcome` ('deleted' → remove; 'resolved' → relocate to the resolved tab).
- Idempotent races (review F2): BOTH apply and dismiss onError reduce 404/400 to
success (comment already gone/resolved), dropping it from the cache instead of
a red error — restores the #315 apply idempotency the ephemeral delete would
otherwise break.
- i18n Dismiss / "Не применять" (ru/en).
Not done (flagged): deleteCommentMark on the normal /comments/delete path — left
out (would change every non-suggestion delete + needs gateway injection; the
interactive client already strips the mark via unsetComment). Out of scope per
the issue.
Tests: server — apply/dismiss delete-vs-resolve fork, all four dismiss state
guards, the deleteCommentMark handler, controller authz (dismiss=canComment,
apply=canEdit), AND a mark-removal-failure test proving the row is NOT deleted +
the error propagates (F1). client — Dismiss show-conditions, outcome cache
reconciliation, and 404 idempotent race for BOTH dismiss and apply (F2).
Verified: server tsc clean; comment+collaboration jest 144 passed. client tsc
clean; vitest 905 passed | 1 expected-fail.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Editorial roles (Corrector/Factchecker) brute-forced `get_node` block-by-block to
find occurrences (unquoted «ё», straight quotes, «т.е.»), burning tokens. New
`search_in_page(pageId, query, {regex?, caseSensitive?, limit?})` reads the page's
ProseMirror JSON via the existing getPageRaw and searches it IN MEMORY — no server
endpoint, no DB/schema change, no touch to the packages/mcp/src/lib schema mirror.
New pure `searchInDoc(doc, query, opts)` (packages/mcp/src/lib/page-search.ts):
recursive descent to each TEXT CONTAINER (paragraph/heading/table-cell paragraph),
glues its inline text via `blockPlainText` (a match survives inline-mark
boundaries — e.g. «т.е.» split across bold/italic), searches literal (indexOf) or
regex, and returns `{ total, truncated, matches:[{ nodeId, blockIndex, type,
before, match, after }] }`. `nodeId` is the container's attrs.id or the
`#<topLevelIndex>` of the enclosing top-level block — the SAME ref format
get_node/patch_node/comment-anchoring accept (verified identical to getNodeByRef),
so the agent goes straight from a hit to a targeted comment; `before`/`after` are
~40-char windows for a unique selection. `total`/`truncated` always reported (never
silent truncation). Lives in the SHARED_TOOL_SPECS registry → exposed in BOTH
transports (external /mcp + in-app AI-chat), with a SERVER_INSTRUCTIONS line and a
DocmostClientLike signature + contract-test entry. Corrector/Factchecker prompts
get a one-line "use search_in_page first" hint (versions bumped, catalog hash lock
refreshed).
Guards: empty/whitespace query → clear error; invalid regex → clear error (not a
generic 500); zero-length regex matches (`\b`, `a*`) skipped with lastIndex
advanced (no loop/flood); MAX_PATTERN_LENGTH=1000, MAX_CONTAINER_TEXT=100k bound
each exec; limit clamped [1,200] (default 50).
Tests: new page-search.test.mjs (17) — literal+regex, case-sensitivity,
mark-boundary glue, nodeId for paragraph/heading (attrs.id) and table-cell
(#<index> fallback), context bounds, limit/total/truncated + clamp, invalid
regex/empty/over-long errors, zero-length skip, empty-doc null-safety.
mcp: tsc clean; node --test 467 passed (+17). apps/server: tsc --noEmit clean
(DocmostClientLike + wiring). catalog check.mjs OK.
Known limitations (from internal review, non-blocking):
- Residual ReDoS: a crafted catastrophic-backtracking pattern (e.g. `(a+)+$`)
against a large single container can hang the event loop — JS regex is not
interruptible, so the length caps bound the base but not the backtracking.
Realistic exposure is low (containers are small; the pattern is supplied by the
authenticated model). Candidate for a follow-up hardening (safe-regex validation
or a worker+timeout) if it matters.
- Case-insensitive LITERAL search folds via toLowerCase; a char whose lowercase
differs in length (e.g. Turkish İ) BEFORE a match could shift the context
window — negligible for the RU/EN editorial scenario.
- On a `#<index>` table-cell fallback, `type` is the inline container ("paragraph")
while nodeId addresses the top-level block — addressing is correct; the field is
documented as the container's type.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The AI agent (MCP + in-app chat) saw ALL comments incl. resolved via two
channels, cluttering its context and breaking fragment search. Default now:
the agent sees only ACTIVE discussions; resolved is opt-in. Active anchors and
threads are always kept.
Channel 1 — resolved comment anchors on agent reads (converter option):
`convertProseMirrorToMarkdown(content, options?)` gains
`options.dropResolvedCommentAnchors` (default false — zero change for every
existing caller incl. git-sync). Both `case "comment"` emitters (top-level and
the raw-HTML inlineToHtml path) emit BARE text (no `<span data-comment-id>`) when
`resolved && the flag`; active anchors keep their wrapper. mcp `getPage` passes
the flag; `export_page_markdown` does NOT (lossless export must preserve resolved
anchors — that is why it is an opt-in option, not unconditional); `get_page_json`
is untouched (lossless PM JSON). Built on the #293 package converter.
Channel 2 — `list_comments` default active-only: `listComments(pageId,
includeResolved=false)` now returns `{ items, resolvedThreadsHidden }` (was a
bare array). By default a RESOLVED top-level thread is hidden wholesale — the
root AND every reply anchored to it (a thread is gated only by its root's
resolvedAt; a resolved reply under an ACTIVE root stays). `resolvedThreadsHidden`
counts hidden threads so the agent knows to re-query. `includeResolved:true`
returns everything. The `includeResolved` param is added to both tool
registrations (MCP index.ts + in-app ai-chat-tools.service.ts); `DocmostClientLike`
signature updated. Server `findPageComments` is NOT touched — the web UI's tabs
depend on the full feed; filtering is only at the mcp-client level. All internal
call sites (export_page_markdown / checkNewComments / transformPage) updated to
`.items` with `includeResolved:true` to keep their full-feed behavior.
The comment model is assumed FLAT (a reply's parentCommentId points at the
thread root) — documented in the filter; a future reply-of-reply model would
need a root-walk there.
Tests: resolved-comment-anchors.test.ts (6 — anchor dropped with flag / kept
without, for BOTH emitters; active always kept); list-comments-resolved.test.mjs
(4 — resolved thread+reply hidden + counter; includeResolved:true returns all;
an ACTIVE thread with a RESOLVED reply is NOT hidden).
package vitest: 664 passed; tsc clean. mcp: node --test 458 passed; tsc clean.
apps/server + git-sync: tsc clean (converter option default-off).
NOTE: based on feat/293-B (#293/#326 STEP 5) — the converter lives in the
package; this PR is stacked on #333 and its base retargets to develop once #333
merges. mcp/build is gitignored (not committed).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Audit of all 41 tool descriptions against the actual implementation found
factually wrong or misleading texts:
- list_comments claimed '(paginated)' — it takes only pageId and returns ALL
comments in one call (internal pagination); now also states that RESOLVED
threads are included and how to filter them. In-app twin synced.
- search claimed the limit default is 'applied by the client' — the client
deliberately omits it so the SERVER applies its default.
- create_page's '(automatically moves it to the correct hierarchy)' said
nothing useful — now documents parentPageId nesting semantics; move_page
drops the stale 'essential for organizing pages created via create_page'.
- share_page now warns the page becomes accessible to ANYONE with the URL.
- get_page (both transports) now explains inline <span data-comment-id> tags
are comment anchors (incl. resolved) — markup, not page text.
- patch_node/delete_node/insert_node pointed only at the expensive page-JSON
view for block ids — now route through the cheap page outline first.
- docmost_transform marks 'Примечания переводчика' as the DEFAULT
notesHeading, overridable for non-Russian pages.
Checks: @docmost/mcp tests 450/450 (incl. the server-instructions guard);
server ai-chat-tools spec 20/20; mcp build/ artifacts rebuilt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Brings the git-sync epic up to current develop (79 commits) so it can land —
prerequisite for the #293 converter unification, which restructures git-sync.
Conflict resolutions:
- translation.json (en+ru): union of #119's git-sync keys and develop's agent
keys; restored a ru-RU key develop had dropped so en/ru stay in parity.
- history-item.tsx: keep develop's AgentAvatarStack (the #300/#319/#320 avatar
rework) AND #119's GitSyncBadge provenance branch; drop the superseded
AiAgentBadge (and its now-unused component/test). Test updated to match.
- collaboration.handler.ts: keep #119's flushPendingStore (QA #119 pre-merge
flush) AND develop's generic withYdocConnection<T> (the #315 closure-capture
version) — git-sync callers pass sync void fns which <T=void> handles.
- comment.ts (editor-ext): take develop's isNodeRuntime SSR guard; it subsumes
#119's isInteractiveBrowser/GS-EXPORT-500 fix (any Node render → static spec).
Build policy: honor #119's approved design — packages/{mcp,git-sync}/build are
gitignored and built in CI/Docker (Dockerfile pnpm build + COPY both), so
develop's committed build/ copies are removed from the index (no more src/build
drift). NOTE: after this merges, rebuild git-sync in any working checkout.
Verified: editor-ext 249/249, collaboration jest 204/204, page-history 13/13,
schema-attribute-contract + converter-gate 32/32; editor-ext/mcp/git-sync tsc
clean. The two residual apps/server tsc lines are the worktree node_modules
symlink resolving @docmost/git-sync to a stale sibling checkout, not a merge
defect (jest green).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- editorial roles (ru/en): proofreader and line editor attach suggestedText
replacements to targeted fixes; fact-checker ALWAYS attaches the ready
correction for [Incorrect] verdicts; structural editor and narrator get a
light-touch rule for in-place rewordings; role versions bumped and the
content-hash lock refreshed
- MCP SERVER_INSTRUCTIONS: route 'propose a concrete text fix for one-click
human approval' to create_comment with suggestedText (unique-selection
reminder); build/ artifacts rebuilt
- AI-chat SAFETY_FRAMEWORK: mention the comment-suggestion capability so the
default assistant offers ready fixes instead of only describing changes
Checks: catalog check.mjs OK; @docmost/mcp tests 448/448; server
ai-chat.prompt spec 28/28.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
F1 [blocking]: a suggestion whose anchor matched via normalization could never
be applied (spurious 409). The comment mark lands on the doc's ACTUAL text
(Docmost auto-converts to typographic quotes/dashes/nbsp), but the stored
selection — used as expectedText at apply — was the raw ASCII agent input
(+substring(0,250)). So replaceYjsMarkedText's strict joined!==expectedText
always failed and threw "text changed" though nobody edited. Fix: new pure
getAnchoredText(doc, selection) reconstructs the exact raw doc substring the mark
covers (slicing identical to spliceCommentMark); on the suggestion path
client.createComment stores THAT as selection, so expectedText equals the marked
text and apply returns applied:true. Live anchoring still uses the raw agent
selection (normalization still finds the anchor). Truncation raised 250->2000
(+ DTO @MaxLength(2000)) so the anchored substring is never cut below the mark
span. Ordinary comments unchanged. AI-chat shares client.createComment, so
covered. Regression tests: getAnchoredText raw-vs-ASCII; create payload selection
is the typographic substring; apply with typographic expectedText -> applied.
F2 [blocking]: added comment.controller.spec.ts pinning that validateCanEdit runs
before applySuggestion (Forbidden -> applySuggestion never called; happy path ->
called; missing comment -> 404 without authorizing).
MCP 448 pass; server comment+yjs 54 pass. MCP build/ rebuilt.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Agents can attach a suggested replacement when creating an inline comment, via
both the MCP create_comment tool and the AI-chat createComment tool.
Because applying a suggestion edits the EXACT anchored text, an ambiguous anchor
would let Apply corrupt the wrong occurrence. So when suggestedText is set the
selection must occur EXACTLY ONCE:
- new countAnchorMatches(doc, selection) counts occurrences across all blocks
(same normalization/traversal as canAnchorInDoc), counting occurrences (2 in
one block => 2) — stricter than block-count, never under-counting distinct
occurrences (false-unique is the dangerous direction).
- client.createComment gains suggestedText: a pre-check (getPageJson +
countAnchorMatches: 0 => not-found, >=2 => ambiguity error) before create, and
an AUTHORITATIVE live check inside the anchoring mutation that recomputes on the
live doc and, if != 1, aborts and rolls back the just-created comment (reusing
the existing safeDeleteComment "anchor not found" path). Ordinary comments keep
first-occurrence behavior unchanged.
- suggestedText is rejected on a reply or without selection in all three layers
(MCP handler, MCP client, AI-chat tool), mirroring the server DTO/service.
- filterComment surfaces suggestedText/suggestionAppliedAt/suggestionAppliedById.
- DocmostClientLike.createComment signature updated. MCP build/ rebuilt.
Tests: countAnchorMatches (0/1/N, within/across/nested block, span nodes,
quote normalization); createComment (ambiguous refused pre-create, reply and
no-selection rejected, unique succeeds and forwards suggestedText, filterComment
surfaces it); ai-chat schema accepts suggestedText. MCP 443 pass; ai-chat 601 pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Server side of agent comment suggestions.
- CreateCommentDto gains optional suggestedText (<=2000). CommentService.create
accepts it ONLY for a top-level inline comment with a non-empty selection,
requires it be non-empty and differ from selection (else BadRequest), and
stores it.
- POST /comments/apply-suggestion (ApplySuggestionDto { commentId }): authorizes
with validateCanEdit (applying edits page text) BEFORE any structural check or
mutation, then CommentService.applySuggestion:
- runs the phase-3 collab event applyCommentSuggestion on `page.<pageId>` to
atomically check-and-replace the marked text, returning { applied, currentText };
- applied → stamp suggestion_applied_at/by, auto-resolve the thread, ws
commentUpdated, audit COMMENT_SUGGESTION_APPLIED;
- already-applied (DB) → idempotent success (no re-apply), self-healing the
resolve if it was missed — satisfies the issue's double-click / two-user
race requirement;
- collab verdict applied:false && currentText===suggestedText → idempotent
success (crash between doc mutation and DB write);
- text changed → 409 ConflictException carrying currentText;
- gateway undefined/throw → hard error, never a silent success.
- audit-events: COMMENT_SUGGESTION_APPLIED.
Tests: create validation (reply/no-selection/equal-to-selection rejected;
valid stored) + applySuggestion verdict branches incl. both idempotent paths.
jest src/core/comment: 33 passed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
POST /api/ai-chat/bound-chat 500'd with Postgres 22P02 because the client
sends a page slugId (10-char nanoid) in the request `pageId` field, which the
server passed straight into the UUID `page_id` column. The chat-to-document
binding silently broke (client fail-softs to a new chat) and every slug-URL
page open logged a 500.
Fix: resolve the incoming id to a real page UUID on the server. PageRepo.findById
already accepts both a uuid and a slugId (isValidUUID→slugId fallback), so
boundChat now resolves the page first, guards it against a foreign/unknown
workspace (returns {chatId:null} before any chat lookup — no cross-workspace
probe), and looks up the latest chat by the resolved page.id (real uuid).
Client: renamed the local pageId→slugId for clarity (the value is a slugId);
the wire body key stays `pageId` so the DTO is unchanged. DTO left @IsString()
(a @IsUUID() would only turn the 500 into a 400 and still break binding).
Test: bound-chat spec asserts a slugId resolves and findLatestByPage is called
with the real uuid; a foreign-workspace page → {chatId:null} without a chat
lookup (no leak); an unknown id → {chatId:null}, no throw.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F1: remove an accidentally-committed self-referential symlink
packages/mcp/node_modules/node_modules -> an absolute build-machine path (leaked a dev
home path, a pnpm artifact useless in the repo), and add a targeted ignore so it can't
recommit.
F2: the commentUpdated broadcast re-emitted the caller's pre-loaded comment mutated in
place, so the {agent,launcher} stack survived only because the controller happened to
load it with includeCreator:true — the fragile coupling that let the stack vanish on
edit once already. update() now RE-FETCHES the enriched comment before broadcasting,
symmetric with create()/resolveComment() (the row is already persisted), so all three
broadcasts carry the stack regardless of any caller's pre-load. Adds a caller-contract
test asserting all three broadcasts emit agent/launcher for an agent comment and neither
for a non-agent one, spotlighting the update path (non-vacuous vs the old re-emit).
F3: add a direct test of the page-history attachPageHistoryAgent mapping (its distinct
lastUpdatedSource/lastUpdatedAiChatId/lastUpdatedBy column set): role / no-role / MCP /
non-agent, and that the internal agentRole join column is stripped.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The same tool metadata (zod schema + model-facing description) was hand-duplicated
between the standalone MCP server and the in-app AI-chat agent, so every tweak had to
land in two places and copies drifted (a materialized parity bug). The shared
transport-agnostic registry (packages/mcp/src/tool-specs.ts) already de-duplicates 14
tools; this migrates two more genuinely-identical ones — patch_node/patchNode and
insert_node/insertNode. The canonical description is a strict SUPERSET of both originals
(keeps MCP's "without resending the whole document" + table-structure/anchor guidance
AND the in-app "reversible via page history" / "exactly one of anchorNodeId or
anchorText" framing — no model-facing guidance dropped); the schema is identical (the
in-app side just gains MCP's .min(1) on ids, a safe tightening). Each transport keeps its
own execute/auth wrapper, and the in-app parseNodeArg node-arg normalization is unchanged.
The three table tools are intentionally NOT merged (a real param-name divergence:
table vs tableRef) — documented on both sides. Other per-transport divergences
(search/share/create_comment/transform/list_pages) are left separate with a short comment
explaining why (the issue asked to flag these as intentional). DocmostClientLike stays a
hand-mirror (the ESM/CJS boundary blocks a compile-time type import; a runtime drift-guard
already pins it). Also fixes a latent contract-spec bug: derive `required` from
`instanceof z.ZodOptional` (matches the emitted JSON schema) instead of `isOptional()`,
which wrongly reported z.any() fields as optional.
Partially addresses #294.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F1: pc.title (untrusted cross-user page title) was interpolated raw into the
markdown export heading. Reusing escapeAttr alone (the prompt sink's XML-attribute
sanitizer, strips < > ") is insufficient here because the sink is MARKDOWN: link
/image syntax survives, so a title like  or [phish](http://evil)
injects a remote image / clickable link into the downloaded .md disguised as a
trusted system annotation. Add markdownHeadingSafe() = escapeAttr() + backslash-
escape [ and ] (disables both [text](url) and ; a bare (url) is inert).
F2: cover the title branch — a title that collapses to empty via escapeAttr falls
to the bare heading (no ("")), and a link/image-injection title is neutralized
(non-vacuous vs the escapeAttr-only version).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The #274 page_changed marker lived only in the ephemeral system prompt, so the
diff the agent saw was invisible in the chat export/history, and the note was
too weak — the agent still overwrote the user's manual edits with a full-page
replace.
- Persist the diff the agent saw as metadata.pageChanged on the assistant row
(flushAssistant), threaded into all five flush call sites in stream(). Model
replay (rowToUiMessage/rowParts) reads only metadata.parts, so the sibling
never re-injects the note into the model context on later turns.
- Render the persisted diff as a labelled block (en/ru) before the message body
in the server-side Markdown export (chat-markdown.util.ts).
- Strengthen PAGE_CHANGED_NOTE: mandate a fresh getPage re-read and targeted
edits (editPageText/patchNode/insertNode/deleteNode) instead of a whole-page
replace, and never revert or overwrite the user's edits.
Tests: prompt, export and service specs updated; 114 pass, tsc clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
F1: escape the collaborative page title before interpolating into
<page_changed page="..."> (and the pre-existing openedPage attr) — strip
<>" and collapse whitespace, so a crafted title can't break out of the
attribute into the system prompt (cross-user injection).
F2: neutralize <page_changed>/</page_changed> occurrences inside the diff body
so a crafted line can't close the block early.
F3: remove the dead content_hash column (written every turn, never read) —
migration, repo, service hashing + crypto import, db.d.ts, spec asserts.
F4: test the best-effort catch branches (detectPageChange / snapshotOpenPage
swallow errors and don't break the turn).
F5: soften the overstated 'diff cannot smuggle instructions' comment to
defense-in-depth framing referencing the F1/F2 mitigations + safety sandwich.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The agent rebuilds context from DB each turn and didn't know the user manually
edited the open page since its last response, so it could overwrite those edits.
Add a per-turn ephemeral <page_changed> note in the system prompt (twin of
INTERRUPT_NOTE, self-clearing) carrying a unified Markdown diff of what changed
since the END of the agent's previous turn.
- New ai_chat_page_snapshots table (migration + hand-declared db.d.ts/entity
types) storing the page Markdown per (chat,page) at each turn's end.
- Pure computePageChange util (whitespace-normalized unified diff via the
existing jsdiff dep, 6KB cap + getPage hint).
- Turn start: if the open page's updatedAt moved past the snapshot, diff current
vs snapshot; non-empty -> PAGE_CHANGED_NOTE in the safety sandwich.
- Turn end: upsert the snapshot on EVERY terminal path (onFinish/onError/onAbort,
once) so the agent's own edits are excluded by construction even on aborted
turns.
All best-effort (never breaks/latency-regresses a turn); fast path when updatedAt
is unchanged. Server-only.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add an inline spoiler (Telegram/Discord-style hidden text): a TipTap mark
`spoiler` rendered as <span data-spoiler="true" class="spoiler">, blurred via
CSS and revealed on click (UI-only is-revealed class, never persisted).
- packages/editor-ext: the Spoiler mark (inclusive:false, set/toggle/unset
commands, ||text|| input rule), exported; a lossless turndown rule emitting
raw inline HTML; round-trip test.
- apps/client: SpoilerView mark-view (ReactMarkViewRenderer, Link pattern),
registration in extensions, bubble-menu toggle button (editable only), CSS
(blur + @media print reveal), en/ru i18n.
- apps/server: register Spoiler in collaboration.util tiptapExtensions so the
mark survives HTML<->JSON export/index/import/Yjs; a test proving the public
share keeps the spoiler (it isn't stripped with comments).
No keyboard shortcut: the proposed Mod-Shift-s collides with Strike (and
Mod-Shift-h with Highlight); the ||text|| input rule + the bubble-menu button
cover ergonomics.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F11: PRE_SEED_TTL_SECONDS 45->120 (= client REINDEX_POLL_CAP_MS). At concurrency
1 a queued reindex can wait past the old 45s; if the pre-seed expired while
pending, getMasked fell back to the COUNT and reported done, so the client
stopped polling and missed the climb. Tie the pre-seed TTL to the client cap.
F12: extend the lockstep integration spec — insertPage takes content; a
text_content=null + text-node-content page is IN and a math-only page is OUT,
pinning the structural "type":"text" clause (and the jsonb space-after-colon).
F13: list all three embeddable clauses in the reindex JSDoc/inline comments.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
guardedFetch blocks loopback/private/link-local/metadata IPs and never calls
fetch; decryptHeaders fails open (returns undefined, warns once, no blob leak).
yjs.util setYjsMark/removeYjsMarkByAttribute/updateYjsMarkAttribute on real
Y.Docs. SHARED_TOOL_SPECS<->in-app parity (name/desc/input-schema; a dropped or
renamed wiring fails). Replace the tautological storage.service spec with
driver-delegation checks across every public method.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Apply review suggestions to the real-files block in
ai-agent-roles-catalog.provider.spec.ts (test-only):
1. Fix inaccurate comment: there are 5 content YAML files (index +
four per-bundle/lang files), not 6.
2. Improve isolation: read/parse the real index lazily inside tests
(via loadRealIndex) instead of in the describe body, so a broken
real file fails only these catalog tests, not collection of the
whole spec (incl. the unrelated mocked-remote provider tests).
3. Add the symmetric slug check: each language file's slug set must
equal the declared slug set (no undeclared/extra roles), matching
scripts/check.mjs's exact two-way correspondence.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The blob-sandbox feature (#243/#250) made AiChatToolsService.forUser()
eagerly call this.sandboxStore.asSink() while wiring the stash tool, but
the spec still passed an empty {} as the sandboxStore constructor arg.
That object has no asSink method, so all 19 tests in the suite failed in
CI with 'TypeError: this.sandboxStore.asSink is not a function'.
Replace the stale {} mock at all 4 constructor sites with a no-op sink
exposing asSink() -> { put, has, evict } (jest.fn()). These tests never
execute the stash tool, so a no-op sink is sufficient for forUser() to
wire successfully. Test-only change; production code is unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Provider tests only exercised synthetic stringifyYaml fixtures, so a
hand-conversion error in one of the 6 real catalog files (index.yaml,
bundles/{editorial,research}/{en,ru}.yaml) — a stray quote/colon in a
description, a broken emoji/arrow, a block-scalar indent slip that
silently changes or drops instructions — was caught by no automated
test. scripts/check.mjs is the only other guard and is wired into no
CI/turbo/husky step.
Add a real-files test block that reads each shipped file off disk,
parses it with the SAME options the provider uses
(strict: true, maxAliasCount: 100), and validates it through the
provider's own exported type guards (isCatalogIndex / isCatalogBundleFile
/ isCatalogRole). It is driven from the real index so new bundles/langs
are auto-covered, asserts the editorial bundle still ships fact-checker,
and requires every declared role to be present with non-empty
instructions/name in each language file.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F1: clear the "Reindex now" spinner once the poll cap fires. Gate the
reindexing part of the button's loading state on the active poll window
(reindexDeadline !== null) so a run that outlives the 120s cap no longer
leaves the button stuck-disabled with a stale `reindexing: true`; the
admin can restart.
F2: rewrite reindexWorkspace JSDoc to describe the EMBEDDABLE page set
(text OR existing embeddings), matching getEmbeddablePageIds /
countEmbeddablePages instead of the old "every non-deleted page".
F3: extract the shared embeddable-content predicate into a private
PageRepo.embeddablePredicate helper, called by both countEmbeddablePages
and getEmbeddablePageIds, removing the verbatim duplication. Behavior is
identical (lockstep int-spec stays green).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Security (must-fix):
- sandbox.controller: the anonymous GET /api/sb/:id response now sets
X-Content-Type-Options: nosniff, a restrictive CSP, and Content-Disposition=
attachment for any mime outside a raster-image allowlist (png/jpeg/gif/webp/
avif). entry.mime is attacker-controlled, so an evil.svg/evil.html could
otherwise execute script inline on the Docmost origin (stored XSS). Mirrors
the public attachment route's hardening.
Stability:
- client.stashPage: reconcile mirrors AFTER the final document put, not only
before it. The doc blob is the newest entry and FIFO eviction drops the
oldest = this stash's own images, so the stored doc could reference an
evicted blob (consumer 404) and over-report images.mirrored. A bounded loop
now reverts doc-put-evicted mirrors, drops the stale doc blob, and re-puts
until stable. Regenerated packages/mcp/build/.
- sandbox.controller: emit Cache-Control on the 304 branch too (ttlSeconds is
computed before the conditional check).
Docs:
- Bump the MCP tool count 39 -> 40 across all READMEs and AGENTS.md (the
registry now exposes exactly 40 tools).
Refactor:
- SandboxStore.asSink() centralizes the {put,has,evict} sink + uri<->id
mapping; the embedded-MCP and in-app agent-tools wiring sites share it.
Tests:
- security headers (inline vs attachment, nosniff, CSP), 304 Cache-Control,
putAndLink URL form, has()/remove(), asSink() round-trip, getSandboxPublicUrl
(trailing-slash trim + APP_URL fallback), and a stash test where the doc put
itself evicts a mirrored image.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Security:
- stash_page: reject path-traversal / percent-encoded srcs before the authed
loopback fetch (resolveInternalFilePath), closing an SSRF/exfiltration hole
where a crafted node.attrs.src could read an arbitrary internal GET endpoint
into the anonymous sandbox.
Stability:
- stash_page: revert + recount mirrors FIFO-evicted by a later put in the same
stash (no dangling sandbox refs, honest images.mirrored/failed); free image
blobs if the final document put throws.
- Reject/clamp non-positive SANDBOX_TTL_MS to the 1h default (warn once).
- Log mirror failures unconditionally (console.warn, no blob bodies).
Cleanup / architecture:
- Remove dead expiresAt from SandboxPutResult.
- Centralize the /api/sb route in SANDBOX_ROUTE_SEGMENT/SANDBOX_API_PATH and
move URL composition into SandboxStore.putAndLink; drop the duplicated sink
closures and the now-unused EnvironmentService injection from McpService and
AiChatToolsService.
- Un-export isInternalFileUrl; document the process-local (instance-bound)
sandbox limitation in the tool description and .env.example.
Docs/tests:
- README/README.ru: 38 -> 39 tools + stash_page entry.
- Add traversal/normalize/recursion unit tests, stash self-eviction +
doc-put-throw + empty/octet-stream mock tests, controller If-None-Match
(wildcard/weak/list) + Cache-Control tests, and SANDBOX_TTL_MS validation
tests. Regenerate packages/mcp/build.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Blocking (review id 2514):
- [security] Forbid symlinks in vaults. ensureServable now sets
core.symlinks=false in each vault's local git config (a pushed symlink is
checked out as a plain file, never a real link), and the engine cycle wraps
every read/write/mkdir in an lstat/realpath guard (new path-guard.ts) that
refuses a path that is — or traverses — a symlink, or whose realpath escapes
the vault root. Prevents a writer from publishing /etc/passwd or the server
.env, or writing outside the vault. Adds unit tests (path-guard.test.ts) +
a read-guard integration test (cycle.test.ts) + real lstat/realpath in the
roundtrip integration test.
- [simplification] Delete dead lib/diff.ts + test/diff.test.ts and drop the
now-unused @fellow/prosemirror-recreate-transform dependency.
- [documentation] Add a CHANGELOG [Unreleased] → Added entry for git-sync.
Warnings:
- [test-coverage] Cover the CREATE-branch conflict-markers guard (a new .md with
markers and no gitmost_id is recorded as a create failure, never created).
Suggestions:
- [stability] Bound each `git config` in ensureServable with a timeout.
- [authz] Trigger endpoint resolves spaceId workspace-scoped and 404s a foreign
space before any vault directory is created.
- [stability] Attribute git-initiated moves to the service account
(lastUpdatedById), via an optional actor param on PageService.movePage.
- [documentation] Document the per-space autoMergeConflicts toggle in AGENTS.md.
- [test-coverage] Cover the unterminated `:::` callout fence fallback.
- [simplification] Move test-only roundtrip-helpers.ts out of src/ into test/.
Architecture:
- Move the Yjs/ProseMirror merge primitives (yjs-body-merge, three-way-merge,
lcs + specs) into collaboration/merge/, breaking the collaboration →
integrations/git-sync dependency cycle this PR introduced.
- Port the schema-surface drift gate to packages/mcp (the mcp schema mirror had
none); pins 52 entries.
Deferred (with rationale in the review thread): the incremental-pull perf
warning (correctness-neutral; needs a high-water-mark design + its own tests on
the data-loss-critical path) and the redis-sync rolling-deploy mixed-version
edge (the deficient behavior is in already-released old-instance code; the new
code is correct on both sides; impact is a transient rollout-window artifact).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add an ephemeral, process-local blob store so the in-app agent (and the
embedded MCP) can hand a large page document and its images to an external
consumer WITHOUT routing the bytes through the model context or Docmost auth.
- SandboxStore (@Injectable singleton): Map<uuid,{buf,mime,sha256,expiresAt}>
in RAM only. put() picks a per-blob cap by mime (image vs doc), enforces a
total-bytes RAM guard with oldest-first eviction, and stamps a TTL; get()
lazily expires. sha256 computed at put() doubles as the strong ETag. An
unref'd sweep interval clears expired entries and is cleared on destroy.
- GET /api/sb/:uuid anonymous controller: serves raw bytes with Content-Type,
Content-Length and ETag=sha256; 404 on missing/expired/non-UUID (anti-
traversal), 304 on a matching If-None-Match. No tokens, no 401 — the
capability is the unguessable UUID + short TTL + TLS. Auth-exempt the same
way as /api/files/public (no JwtAuthGuard) plus an /api/sb entry in main.ts's
workspace-resolution preHandler so a remote consumer with no workspace host
is not rejected.
- stash_page tool in both layers (MCP resource_link + in-app {uri,size,sha256,
images}). client.stashPage serializes the get_page_json shape, mirrors every
INTERNAL file/image src (type-agnostic, covers drawio/excalidraw/video/file)
into the sandbox under Docmost auth and rewrites src to the sandbox URL;
external http(s) srcs are left untouched; dedup by src; a failed image fetch
is counted, never aborts the doc.
- SANDBOX_PUBLIC_URL / SANDBOX_TTL_MS / SANDBOX_MAX_BYTES /
SANDBOX_MAX_IMAGE_BYTES / SANDBOX_MAX_TOTAL_BYTES wired through the
environment service + validation + .env.example.
- SandboxModule (@Global) provides the shared store to the controller,
McpService and AiChatToolsService (same instance for put and get).
Tests: SandboxStore (round-trip, sha256, TTL lazy + sweep, caps, eviction),
SandboxController (200+ETag+CT+CL, 404 missing/expired/non-UUID, 304), and a
mock-HTTP stashPage test (mirror+rewrite internal, keep external, dedup, failed
image counted, returns only a link). Interoperates with the vvzvlad/habr-mcp
consumer's anonymous-GET + sha256-ETag + resource_link contract.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
1. gitRemote is NOT yet consumed (the vendored engine has no remote-push path,
SPEC §7). Corrected the buildSettings docstring (it wrongly called gitRemote
"load-bearing") and marked the env -> validation -> getter -> buildSettings
chain as inert SCAFFOLDING for the deferred remote-push feature at all three
sites. Kept the wiring (harmless; removing only churns).
2. .env.example: document that GIT_SYNC_REMOTE_TEMPLATE substitutes the literal
"{spaceId}" per-space (with the example), so an operator doesn't point every
space at one remote.
3. Extracted the copy-pasted CJS->ESM dynamic-import bridge
(`new Function('s','return import(s)')`) into one shared
common/helpers/esm-import.ts; git-sync.loader, docmost-client.loader and
mcp.service now import it and keep their own typed loadX() wrappers.
Deferred (notes only, not implemented):
- lcs.ts + three-way-merge.ts could move into packages/git-sync, but that engine
is vendored (manual re-sync) — added a one-line note at three-way-merge.ts to
revisit once the re-sync story is settled.
- schema-core single source + BullMQ/fencing remain documented from prior rounds.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
After rebasing onto develop, movePage runs its cycle-check + UPDATE inside
executeTx(this.db) (develop #207 advisory-lock/atomic cycle-guard). The
git-sync provenance specs still passed a bare `{}` db, so executeTx hit
`db.transaction is not a function`. Reuse the same trxStub Proxy + transaction
mock the develop movePage specs use so both the advisory-lock `sql.execute(trx)`
and updatePage resolve. Production movePage keeps BOTH develop's lock/cycle
guard AND git-sync's provenance stamping; this only updates the test harness.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Red-team #13 (conflict markers reaching Docmost) is now a per-space policy
exposed as a UI toggle, instead of a hardcoded behavior. New boolean
`gitSync.autoMergeConflicts` (default FALSE), mirroring the existing per-space
`gitSync.enabled` flag end-to-end (jsonb space settings -> update-space DTO ->
space.service -> client types -> space settings form switch):
- OFF (default, safe): a page whose committed body still has unresolved git
conflict markers is NOT pushed — it is recorded as a per-page push FAILURE
("unresolved conflict markers — resolve in git first"). Recording a failure
(not a soft skip) deliberately HOLDS refs/docmost/last-pushed so the conflict
commit is never marked pushed and a later pull cannot clobber the user's
in-progress resolution; the page retries until the conflict is resolved in git.
- ON: the marker lines are stripped and both sides' content is pushed (the prior
behavior), so the conflict becomes visible/fixable inside Docmost.
The engine Settings carries `autoMergeConflicts`; runPush threads it into the
update AND create paths. The orchestrator's buildSettings reads the per-space
flag from jsonb (strict opt-in like `enabled`, default false).
Tests: redteam-push-cycle #13 rewritten (default -> not pushed + failure + refs
held; ON -> strip-and-push); space.service + edit-space-form + orchestrator
specs extended. git-sync vitest 618, server jest space+git-sync 163, client
edit-space-form 11, server/client tsc clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
agent_coder
claude code agent 227
vvzvlad
agent_vscode
agent_coder
a