Why: Independent review found the cross-read toggle, the project inbox
scope and the archive-to-project action still reachable from an open
scratch (no-project) conversation.
Evidence: npx vitest run (685 tests) and npx tsc -b --noEmit pass locally
against the UI fix branch (SELF); CI on the PR head is authoritative.
Impact: Scratch conversations hide the cross-read setting, always show
the conversation inbox and omit onArchive; real projects are unchanged.
Refs: Agent-System-UI#60; agent-exec receipt u3e-run (cursor, success)
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Accounts without project.create need isolated general-agent chats.
The sidebar groups every scratch session under one "无项目对话" entry
instead of the project switcher, and the client no longer auto-creates
a default project.
Evidence: npx vitest run (64 files, 677 tests) and npx tsc -b --noEmit
pass locally (SELF); CI on the PR head is authoritative.
Impact: createTask uses POST /api/scratch-sessions in scratch mode or
when no real project exists; scratch rows hide cross-read toggles and
project settings; "新建项目…" stays gated by project.create.
Refs: Agent-System-Dev docs/superpowers/specs/2026-10-22-scratch-session-general-agent-design.md;
agent-exec receipt u3-run5 (cursor, success)
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: cursor 3/3 exhausted; PI direct dispatched as host integration work.
Scope: 8 test files (7 modified + 1 new), 255 insertions, 29 new vitest tests covering 11 U4 frontend-auto-discover scenarios end-to-end:
- AutoDiscover (3 new in store.test.ts)
- CapabilityGroup (3 new in rolePermissions.test.ts + 6 new in panels)
- CreatorOwnerMemberDuties (4 new across adminRdmSections/projectmeta/AccountUsersPanel)
- MemberManagementEntryHidden (3 new in adminRdmSections/projectmeta)
- EmptyListNoCreate (4 new in projectmeta/TaskSidebar)
- DefaultProjectDisabled (1 new in store.test.ts/projectmeta)
- RefreshBackoff + RefreshSingleFlight + RefreshHiddenPause + RefreshReconnectCompensate
- ProjectLifecycleConversation (5 new in test/conversation-work.test.ts)
Verification: npm run build ok (5.35s); ./node_modules/.bin/vitest run 656/656 (was 627, +29 new); new 29 tests across all 11 patterns all pass; no production files touched.
Refs: plan revision 3 U5-Frontend; /tmp/iam-link-u5-frontend-pi-verify-20261019.json exit=0.
Why: PR #106 is behind main by 15 commits (password login, regulation
citations, RDM member search) and conflicted in RdmAuditWorkspace.tsx and
rdmManagement.ts; the user authorized merging the renderer line on 2026-10-18.
Evidence: content is the agent-exec transport ec5012fe (cursor worker, unit
frontend-task11-sync-main-50ec836-r2, receipt
/tmp/cleanup-20261018/render/frontend-sync-run2.json, SELF): tsc -b, vitest
541/541 and vite build passed; host reviewed both conflict resolutions as unions.
Impact: no behaviour removed from either side; regulation tab and document
open flow coexist in the materials tab.
Refs: PR #106, Frontend main 50ec836
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: RDM now anchors annotations on nodeId + textStart + textEnd + textHash
(spec 06 §2) and acceptance 10 §5 requires accurate comment ranges. The
workspace only sent whole-node annotations and ignored quote/needsReanchor.
- Selecting text inside one node's title in the reader view records a
code-point range (Array.from indices; surrogate pairs and ZWJ emoji are
counted as runes, half-pair endpoints widen) and sends textStart/textEnd
plus quote, so RDM rejects rather than misplaces a drifted range.
- No selection keeps whole-node behavior; replies never carry range params;
textHash/range are never sent; a range is dropped if it belongs to another
node.
- Selections spanning several nodes or containing no node text record no
range and show a notice that the whole node will be annotated; the composer
always shows the scope with a "whole node" escape.
- Annotation items show the server quote and a visible needs-reanchor state;
review_anchor_invalid: errors map to readable Chinese text.
- Reader titles become role=button spans because browsers do not allow text
selection inside <button>.
Evidence: npm test 40 files / 486 tests pass (30 new in
test/review-annotation-anchor.test.ts); npm run build exit 0;
node scripts/check-knip.mjs "50 <= 51"; design-guard 7/7; playwright
--list compiles 13 tests; headless Chromium check of the selection reader
(single, multi-node, triple-click clamp, emoji offsets).
Impact: document review UI only; base-gate decision source anchoring is
unchanged. Offsets are computed on the displayed node title (no frozen-text
read endpoint exists), guarded by quote equality on the server.
Refs: unified renderer plan Task 13; phase1 F2; spec 06 §2
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: RDM be80ad4 (docs/design/render-node-id-contract.md) accepts workspace tree
node ids (dn-/sec-/legacy-) for review annotate/revise, adds treeNodeId to receipts,
rdm.render.review.view items and trace entries while nodeId stays canonical, points
anchored-to at the tree node id, and marks audit-reader trace entries with
contentMasked=true. Grouping review items by the canonical nodeId left them all
unlocated, and the workspace showed no render trace at all.
Evidence: npx tsc --noEmit exit 0; npx vitest run 39 files / 456 tests passed (12 new
in test/document-render-workspace.test.ts: treeNodeId grouping with nodeId fallback
and orphan rules for annotations and revisions, data-locate targets, annotate/revise
payload carrying the selected tree node id, layer=render filtering, anchored-to
against tree nodes without canonical fallback, masked entries never reading body
fields, empty state); npm run build ok; node scripts/check-knip.mjs 51 <= 51;
e2e/document-renderer.playwright.ts type-checks (tsc exit 0) and
npx playwright test --list loads 13 tests.
Impact: locateNodeId (treeNodeId first, nodeId fallback, otherwise unlocated) is
shared by diagnostics, review grouping and trace. Clicking an annotation or revision
group, a reply, or a trace node locates and re-scrolls like diagnostics. Annotate and
revise send the selected tree node id as nodeId. A 渲染链条 panel reads
rdm.inspect.relations (kind=doc) through the single guarded resource owner (stale
responses dropped) and reloads after preview, submit, export, review and decision
actions. Masked entries show metadata only with a 内容已遮蔽 marker and a panel
notice. The e2e spec is unchanged: §11-5b already asserts that the group title is the
selected tree node, and the leak scan is untouched. Not run against a live backend
or browser here.
Refs: RDM be80ad4 docs/design/render-node-id-contract.md; unified renderer spec 10
§11 items 4, 5 and 7; DESIGN.md §8, §9.2, §11.
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Browser acceptance (/tmp/renderer-browser-acceptance/20261017T061021Z) found
§11-4 showed 25/25 diagnostics as 未定位节点 because the renderer nodeId is the
canonical mapping id (pp.cover.document-number.value) while the workspace tree uses
dn-… ids; RDM now adds treeNodeId to every diagnostic. §11-3 failed only because
the e2e spec still expected 已验证/未发布 although all 44 DHF profiles are published.
Evidence: npx tsc --noEmit exit 0; npx vitest run 39 files / 445 tests passed (4 new
in test/document-render-workspace.test.ts covering treeNodeId, nodeId fallback,
unknown ids and per-node counts); npm run build ok; node scripts/check-knip.mjs
unused exports 51 <= 51; e2e/document-renderer.playwright.ts type-checks (tsc exit 0)
and playwright test --list loads 13 tests. Real rdm.render.profiles payload in
evidence/network/item3-author.jsonl: template rule state=published,
publicationAllowed=true, no blockedReason, refusal controlled_requires_exact_profile.
Impact: normalizeDiagnostics keeps treeNodeId; locateDiagnostic resolves treeNodeId
first and falls back to nodeId, otherwise the diagnostic stays explicitly unlocated.
Badges and the 有诊断问题 filter count by the located tree node. Clicking a located
diagnostic selects the node and scrolls it into view in canvas or reader, also on a
repeated click. Diagnostics loading keeps its single request guard. §11-3 now asserts
已发布, 可正式交付, an enabled option and no reason, and still asserts source, version,
the no-fallback notice and refusal. Not run against a live backend or browser here.
Refs: unified renderer plan Task 13 §11 items 3 and 4; spec 10 §11 item 3;
summary.md root causes §A and §B.
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: The base gate (cmdbus.RequiresProjectWorkDecision) refuses rdm.review.doc.approve/reject/final unless the request carries a committed conversation source as top-level conversationId and anchorMessageId. The workspace sent bare params, so the decision buttons were refused in the real product.
Evidence: vitest 39 files / 441 tests passed (22 new in test/document-decision.test.ts); npx tsc --noEmit clean; npm run build ok; check-knip unused exports 51 <= 51; e2e/document-renderer.playwright.ts type-checks and lists 13 tests. Request shape from pins/agent-system-backend: internal/server/server.go:1276-1304 (body decode), internal/cmdbus/cmdbus.go:109-134 and 385, internal/app/app.go:646-665 (preflight), server.go:1010-1150 (message POST returns the committed message).
Impact: approve/reject/final first post a Chinese decision message to the host project's 文档审签记录 conversation (reused, or created), then dispatch with its id as the anchor (entrance button). Identity is never sent. Any failing step shows a readable refusal and nothing is dispatched. Double-click is ignored while an action is in flight, and retries reuse the operation id so the server returns the same source message. rdmCommand takes an optional source; rdm.render.* is unchanged. The e2e spec drives §11-5c through the workspace only and asserts the anchored, committed receipt. Not run against a live backend or browser here.
Refs: unified renderer plan Task 11; RDM cmdbus gate (pins/agent-system-backend internal/cmdbus/cmdbus.go:109-134); live-verify sequence /tmp/live-verify/run.sh lines 56-70 and 140-146.
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Task 11 requires the document render workspace to offer approve, reject and final decisions, rule bind/unbind for document, document type and project default, and the full delivery package, so reviewers and settings holders no longer need other screens.
Evidence: vitest 38 files / 419 tests passed (new test/document-render-actions.test.ts covers decision entries from frozen duties, decision payloads without identity fields, package availability, refusal codes, binding candidates and payloads, perm gating and SSR visibility); tsc --noEmit clean; knip unused exports 51 <= baseline 51.
Impact: decision entries show only when the command is enabled and the frozen submission lists the session user; binding needs rdm.perm.check project.settings allowed=true; package appears only for ready deliveries; organization-default binding is not offered from the workspace (domain admin only).
Refs: unified renderer plan Task 11, 05 section 9, 06 section 6, 07 sections 1 and 5; RDM plugins/rdm/plugin.go review.doc.* and render.profile.* schemas (7169944, 4b7ac2a, 75ba50c).
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Running §11 against RDM d63f1db showed three Frontend defects. A
validated template rule with a server blockedReason was disabled for
preview, although the server only restricts formal delivery. A reviewer
opening the document in another session could not find the author's
submission until an export existed. Template-instance blocks without a
title (cover, sign-off table, fields) rendered as blank canvas nodes and
reader headings.
Evidence: npx vitest run (35 files, 363 tests passed); npm run build ok;
node scripts/check-knip.mjs (43 <= 43); browser run under
/tmp/renderer-browser-acceptance/20261016T064052Z (item1 172 nodes,
preview generated with the validated template rule).
Impact: Preview is offered for validated/draft rules while formal submit
stays blocked; the current submission comes from rdm.review.doc.status;
untitled nodes show "<kind>（<templateBlockId>）".
Refs: unified renderer plan Task 13 §11
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Independent review of a81e94a..85428b4 found five contract mismatches
against the RDM server: profile sources document-type/project-default were
dropped, the refusal object {code,message} was ignored, submit receipts were
read as `id` instead of `submissionId`, diagnostics were expected inside
preview/export receipts (they live in a separate diagnostics delivery blob),
and rdm.inspect.tree was treated as recursive although it returns one level.
Evidence: New tests use the server JSON shapes (RenderProfileOptions,
RenderDeliveryReceipt, DocSubmissionRec, rdmRenderDownload, InspectTree); 26
of them fail against the previous implementation. npx vitest run: 35 files,
359 tests passed; npm run build: ok; scripts/check-knip.mjs: unused exports
43 <= 43; git diff --cached --check: clean.
Impact: Rules panel lists all five sources with Chinese labels and keeps the
server's selected default; refusals render readably; a new submission is
auto-selected; diagnostics are fetched via rdm.render.download kind=diagnostics
under a sequence/epoch guard with readable failures and a Chinese receipt
state label; the document tree is walked breadth-first (depth 12, 2000 nodes,
concurrency 4) with a truncation notice. Playwright spec asserts per-node
diagnostic anchors after preview (authored, not run).
Refs: unified renderer plan Task 11 review findings
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: unified renderer Task 11 needs a reader/delivery UI over the RDM
document tree: four-level rule selection with controlled-document
no-fallback explanation, preview with node-linked diagnostics, a
structure canvas and a linear reader sharing one selection and filter,
review annotations/revisions, and per-delivery downloads.
Evidence: npx vitest run -> 35 files, 338 tests passed (45 new in
test/document-render-workspace.test.ts, design-guard green);
npm run build -> tsc -b + vite build ok; npx knip -> findings identical
to the pre-existing baseline (check-knip 43 <= 43); git diff --check
clean. e2e/document-renderer.playwright.ts authored, not executed.
Impact: new src/documentRenderPresentation.ts (pure helpers) and
src/components/DocumentRenderWorkspace.tsx, opened from a per-document
"阅读与交付" action in the audit materials tree; store gains an
rdmRender capability probe (fail-closed); audit export reuses the shared
base64 file saver; token-only styles appended to styles.css. Async
state is owned per hook with request guards plus identity epoch, and
the workspace remounts per identity/project/document scope.
Refs: unified renderer plan Task 11
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: a lost write response was keeping the old duty version, so the next save replayed it, and a busy save could stay locked if another request advanced the sequence.
Evidence: SELF vitest duty-draft 6 passed, tsc passed, knip 43, Playwright duty flows 3 passed.
Impact: an unknown save reloads server roles, confirmation resets, and busy clears even if a later read is discarded.
Refs: docs/proposals/2026-10-15-rdm-duty-matrix-rework.md
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: project settings must keep the shared duty draft and must not regain template document creation from main.
Evidence: SELF vitest 292 passed, tsc --noEmit passed, knip unused exports 43.
Impact: the duties tab stays assignment-only while conversation ordering from main is included.
Refs: docs/proposals/2026-10-15-rdm-duty-matrix-rework.md
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Completion or progress from an earlier task could appear in the process of an unrelated newer answer.
Evidence: The new ownership regression fails before the guard and passes afterward; the frontend suite, TypeScript and unchanged export gate pass.
Impact: Project child progress only into its originating live turn while still reconciling background results.
Refs: feature/workbench-flow-completion chronological display requirement
Why: Full execution history must retain outputs after the 500-event API page boundary.
Evidence: Ordered 502-event replay and nonadvancing-cursor rejection tests pass.
Impact: Add regression coverage only; production source is unchanged.
Refs: feature/workbench-flow-completion review testing gap
Why: A rejected result-reference request could discard the user text or uploads after the composer cleared.
Evidence: 286 frontend tests, TypeScript and the unchanged unused-export gate pass; mixed-input browser regression preserves the draft.
Impact: Validate incompatible inputs before clearing, retain failed pending input with its request identity, and require explicit retry while other messages can continue.
Refs: feature/workbench-flow-completion review frontend-races finding
Why: the settings page was mixing duty assignment with whole-document review and chapter objects. The three views need one draft and one save check.
Evidence: SELF vitest run of the frontend suite, 278 tests passed, including duty-draft and design-guard.
Impact: the duties tab assigns existing documents only. Submit, approve, reject, and final stay out of this page. Commands that are not ready keep the tab hidden.
Refs: feature/rdm-duty-matrix
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Repeated progress and coupled scrolling obscured chat, and task choices lacked a reliable independent interaction path.
Evidence: 285 frontend unit tests and 23 browser scenarios pass locally before the shared event-loader extraction; final combination checks follow in Dev.
Impact: Coalesce cumulative progress, retain execution records, show timestamps and quiet notifications, support versioned work actions and explicit result references without automatic chat backfill.
Refs: feature/workbench-flow-completion
Why: the duties tab only listed existing documents — rdm.docs.instantiate
had no UI entry, so a fresh project showed an empty duties tab with no
setup path (owner trial feedback 2026-10-15).
Evidence: new docs-instantiate-params tests green; full vitest run shows
only the 12 pre-existing conversation-work failures also present at clean
HEAD; design-guard green.
Impact: duties tab gains a template picker + title input wired to
rdm.docs.instantiate (backend command rdm.docs.templates lists published
identities); instantiation refreshes the catalog and selects the new doc.
Refs: Agent-System-RDM docs/records/dhf/u4-header-footer-slots.md
Why: Independent review found component-local color derivation, a private transition, and missing dark/molv coverage in the visual candidate.
Evidence: Frontend design guard passes and 19 Playwright scenarios cover light, dark, molv, and 320-1920px layouts.
Impact: Keeps the approved visual direction while moving color ownership to @ui tokens and removing private motion.
Refs: Agent-System-Dev feature/workbench-visual-redesign
Why: Status words shared with a final reply could suppress a child completion record.
Evidence: 272 frontend unit tests, build, and 15 isolated browser scenarios.
Impact: Tag client-side collaboration replay so partial-delta deduplication preserves agent notes.
Refs: Agent-System-Dev feature/work-companion-visibility UI1
Why: A dropped delta batch can leave a short event prefix next to the complete saved message.
Evidence: 271 frontend unit tests, build, and 14 isolated browser scenarios.
Impact: Suppress overlapping partial replay text and thinking while retaining saved content.
Refs: Agent-System-Dev feature/work-companion-visibility UI1
Why: Event replay could hide saved retry notes and leave completed cards outside the active work filter.
Evidence: 270 frontend unit tests, build, and 14 isolated browser scenarios.
Impact: Merge replay into saved history, retain child progress, and locate cards under the right scope.
Refs: Agent-System-Dev feature/work-companion-visibility UI1 UI2
Why: Fast and slow channel cards duplicate across panes and completed replies hide process output.
Evidence: npm test; npm run build; 12 isolated Playwright scenarios with reload and responsive widths.
Impact: Preserve ordered folded process history, load durable run events, and give work cards one detailed pane.
Refs: Agent-System-Dev feature/work-companion-visibility UI1 UI2 UI3
Why: Complete the authorized export entry while preserving current main project and session behavior.
Evidence: SELF full frontend tests and build, 263 unit tests; deferred browser races reproduced then all 6 passed. Raw records stay outside source until Dev task closure.
Impact: Export callbacks are invalidated on project, identity and unmount changes. Command readiness and current main security paths remain intact.
Refs: Dev PR #488; Frontend PR #73; user takeover authorization on 2026-10-14
Why:
Project readiness is a write gate. Applying it to member, permission, and
document status queries hides history that the backend intentionally keeps
readable after archival or closure.
Evidence:
SELF archived and closed read regressions failed before the fix. Full npm
test, npm run build, and git diff --check passed with the amended code.
Receipts: /tmp/project-member-frontend-archived-read-red-20261013.json and
/tmp/project-member-frontend-archived-read-green-20261013.json.
Impact:
Resolve canonical project IDs separately from command readiness. Preserve
known read commands and existing settings tabs for historical projects;
keep mutations and new work blocked with existing disabled controls.
Server authorization remains authoritative.
Refs:
Dev project-member-access Task 4; independent Frontend review P2
Why:
An identity can change after an operation has passed its first response guard.
Later project loads, cleanup calls, and attachment refreshes must not reopen
private context or return the previous account's result to the new session.
Evidence:
SELF regressions failed before the guards: delayed project opening, task
creation success, completed-card cleanup, and upload artifact-ID return.
SELF npm test, npm run build, and git diff --check passed; full command
receipt: /tmp/project-member-frontend-final-checks-20261013.json.
Impact:
Guard post-await continuations with the same identity epoch. Preserve the
original server-side result and ignore stale client projections; do not undo
the previous account's already completed work.
Refs:
Dev feature-project-member-access implementation plan Task 4
Why:
Project members need RDM commands to address the registered domain project.
Failed project states and responses from a previous account must not create
work, report a rejected mutation as successful, or restore private content.
Evidence:
SELF npm test: 32 files and 251 tests passed, including project command,
mutation rejection, archived history, and delayed account response coverage.
SELF npm run build and git diff --check passed. Original regressions were
observed before implementation; root owns composition and independent replay.
Impact:
Keep host IDs for conversation and base settings APIs; send the decorated
domain ID to RDM project commands. Reuse existing sidebar, settings and toast
states. Reset identity projections and invalidate old asynchronous responses.
No deployment, database change, push, merge, or workspace cleanup.
Refs:
Dev feature-project-member-access task card and implementation plan Task 4
Why: The 2026-10-13 login contract makes username/password the default
login, replaces the browser ask_ tab with a one-time recovery key
entry, requires the forced set-password gate to drop the current
password field, and routes admin password recovery through admin-issued
login keys instead of direct password resets.
Evidence: vitest 218 passed, npm run build OK, check-knip 43<=43,
design-guard green (SELF, local run in task worktree).
Impact: LoginPage gains a redeem-only key tab (no key persistence);
ChangePasswordGate shows server-confirmed owner plus new/confirm only
and returns to the login page after success (reloginRequired); App
gates on sessionScope=set_password_only; UsersPanel replaces reset
password with issue/revoke login key UI showing plaintext once.
Refs: Dev task feature/svc-login-recovery-flow; backend companion
commit 467718a.
Why: when rdm.role.list is rejected (e.g. the operator lost the RDM
domain-admin ledger to an acceptance bootstrap), the roles table spun on
'加载中…' forever and the banner dumped the raw 400 receipt JSON.
Evidence: vitest test/rdmroles-reject.test.ts (SELF, run below); manual
repro on 8788 dev 2026-10-13 before the ledger fix.
Impact: placeholder row comes from rolesListPlaceholder (failure row when
a fetch rejected), domain-admin denials map via describeRoleListError to
a readable message naming the rdm.perm.transfer channel; other errors
pass through. No style tokens touched (design-guard clean).
Refs: Dev task feature/rdm-admin-lockout-recovery (R3)
Why: Ordinary explanations were folded with tool activity and the oldest result was selected as current work.
Evidence: Findings #5/#6 red-to-green actual-component and live-projection checks are retained; current full Frontend suite passes 210 cases.
Impact: Preserve business text, errors and file references in reading order; retain active/pinned priorities and newest-first host ordering.
Refs: ce-code-review 20261012-165912-cb241b46 #5/#6, UI03/UI05
Why: Background frames blocked terminal reconciliation, stale failures left confirmations busy, and delayed page hydration changed mobile navigation.
Evidence: Review findings #1/#2/#3 have recorded red-to-green checks; all 210 Frontend tests, tsc and knip pass, plus six real navigation browser checks.
Impact: Keep sequence high-water and operation ownership, merge observations by object, and use an explicit preview-opening intent.
Refs: ce-code-review 20261012-165912-cb241b46 #1/#2/#3, UI04/UI09/UI10/UI11
Why: make material, background work, confirmations and files accessible to nontechnical users across desktop and mobile.
Evidence: Dev SELF receipts for Frontend tsc, 190 tests, knip baseline, build, nine fixture browser cases and two real HTTP/browser smoke cases using MockRunner.
Impact: existing API and permissions remain authoritative; stale events/queries and duplicate effects are guarded, while one file preview and reading state are preserved.
Refs: work-companion-ui design r0.2 and plan r1 U2-U5; UI01-UI13.
Why: auditors need to generate and download a watermarked report from the overview, and that must not appear when the export commands are off or change whether the workspace itself opens.
Evidence: SELF npm test, 138 passed. Playwright download case is in e2e/rdm-audit.playwright.ts and runs under scripts/e2e.sh.
Impact: the overview shows type, generate, draft confirmation, watermark, sha256, expiry, and download only when the three export commands are enabled.
Refs: RDM pull request 301
Co-authored-by: Cursor <cursoragent@cursor.com>
weichenandCursor