Why: Settings plaintext loses to secret_keys, so replacing the env key did not change the key a conversation actually sent.
Evidence: python3 -m unittest tests.test_svc.SvcDbRecoveryTests.test_seed_replaces_old_system_secret tests.test_svc.SvcDbRecoveryTests.test_seed_warns_when_old_secret_stays_active tests.test_svc.SvcDbRecoveryTests.test_seed_puts_without_credentials_when_auth_off (SELF).
Impact: Embedded databases now compare the active system secret fingerprint after seeding, and the backend gitlink carries the adoption fix.
Refs: fix/model-key-secret-overlay
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: establish the authorized task workspace and its initial baseline.
Evidence: preserved precheck receipt in .task/fix-model-key-secret-overlay.md; R0=919f75cf7b39ca357e212a08a0ba5ba3d9a79b51.
Impact: task metadata only; the feature workspace starts at 919f75cf7b.
Refs: .task/fix-model-key-secret-overlay.md
Why: Keep the final PR, component references and operational acceptance boundaries in the authoritative task record.
Evidence: Existing fixed-product review and local verification receipts are linked without relabeling their provenance.
Impact: Task-record metadata only; the reviewed product combination is unchanged.
Refs: Dev #565
Why: Deliver the completed conversation and work-card experience to 8788 after exact-head gates.
Evidence: Independent review and both focused addenda are complete; local test, build and environment-conflict records are retained honestly.
Impact: Pin the final frontend ownership guard and record delivery boundaries; no further product scope is added.
Refs: feature/workbench-flow-completion
Why: Complete execution history must retain outputs beyond one API event page.
Evidence: Independent replay passed both pagination-boundary and stalled-cursor tests; production frontend source is unchanged.
Impact: Pin a test-only frontend follow-up to the reviewed workbench combination.
Refs: feature/workbench-flow-completion review pagination addendum
Why: Track background execution before starting it and preserve generated content and ordinary questions across failure paths.
Evidence: Local full combination test passed before this bounded closure; new fault and race regressions pass, with independent closure and final gates pending.
Impact: Update only the reviewed Backend/UI closure and saved local verification output.
Refs: feature/workbench-flow-completion
Why: New questions, rejected messages and delayed result projection must remain recoverable.
Evidence: Focused review regressions, frontend/UI suites and browser mixed-input checks pass locally; full combination retest and independent closure follow.
Impact: Preserve user intent, repair saved result projections and keep archived work read-only.
Refs: feature/workbench-flow-completion
Why: include the current integration baseline before verification.
Evidence: fetched origin/main=37a8791cb9f736a7f58e87a88389a55095f06e0d.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
Why: Complete the approved conversation and background-work experience without changing the accepted visual theme.
Evidence: Component regressions and 23 local browser journeys pass; review and final exact-head gates remain pending.
Impact: Combine full results, optional decisions, saved process and independent chronology with compatible optional contracts.
Refs: feature/workbench-flow-completion
Why: Archive the merged four-model vision task card onto main.
Evidence: CI run 3307 attempt 2 jobs 5093, 5094, and 5095 succeeded on 5726fb6.
Impact: Task card moves to .task/archive; product tree is unchanged.
Refs: #564
Why: Freeze a shared contract for independent implementation units.
Evidence: make contracts generated both TypeScript consumers.
Impact: Component-only baseline for parallel scoped work.
Refs: feature/workbench-flow-completion
Why: Preserve fixes already merged by parallel work.
Evidence: Component branches were created from freshly fetched origin/main.
Impact: Record the actual shared baseline before generating contracts.
Refs: feature/workbench-flow-completion
Why: Complete the uploaded-material work flow and preserve current backend fixes.
Evidence: Imported READY startup receipt is attached to the task card.
Impact: Establish isolated implementation scope and component baseline.
Refs: feature/workbench-flow-completion
Why: establish the authorized task workspace and its initial baseline.
Evidence: preserved precheck receipt in .task/feature-workbench-flow-completion.md; R0=4ba5c40a609dc07deb34d05761b00c29b44f2355.
Impact: task metadata only; the feature workspace starts at 4ba5c40a60.
Refs: .task/feature-workbench-flow-completion.md
Why: branches merged into main had left their task cards open; backfill status and archive to prevent bookkeeping debt.
Evidence: feature-four-model-vision.md(PR#563) — merges verified via git log --merges plus Gitea API squash lookup against origin/main.
Impact: .task/*.md moved to .task/archive/ with status backfilled; no code or submodule change.
Refs: task.sh close (card-autoclose-r1)
Why: land four-model routing and current-turn image delivery.
Evidence: test, test-pg, and svc-realcombo success on f938b3ca97, run 3303.
Impact: main serves images on capable models; 8788 is not restarted.
Refs: #563
Why: the merge candidate needs a reviewer, base, and head before it can be pushed for CI.
Evidence: independent read-only review of 58e4e14 found no correctness or security issues; b79b4d9 has no non-card diff; task-sync changed=0 against ff3a562, receipt ae2c8f553111ec58179197b85971f25073a98bb948437c9a1ff96ea41adebfea.
Impact: task metadata only.
Refs: .task/feature-four-model-vision.md
Why: include the current integration baseline before verification.
Evidence: fetched origin/main=ff3a562e50375ed592126c36ed2c76327517688a.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
Why: the merge onto current main changed the candidate, so the earlier local test no longer covers this head.
Evidence: SELF make test exit 0; receipt 8225fe2d744ba24554bd9af6d6faf08585e34f89beeb6a6decd97190339a410e; before=58e4e1498bbace37e4937a874430e342e6ff7d27.
Impact: task metadata only.
Refs: .task/feature-four-model-vision.md
Why: the plugin compose task card stayed open after its feature merged.
Evidence: Actions run 3300 jobs 5075 test, 5076 test-pg, 5077 svc-realcombo success on 4542cd7d37.
Impact: archive the card as completed. No product code change.
Refs: #562#558
Why: branches merged into main had left their task cards open; backfill status and archive to prevent bookkeeping debt.
Evidence: feature-svc-plugin-compose.md(PR#558) — merges verified via git log --merges plus Gitea API squash lookup against origin/main.
Impact: .task/*.md moved to .task/archive/ with status backfilled; no code or submodule change.
Refs: task.sh close (card-autoclose-r1)
Why: include the current integration baseline before verification.
Evidence: fetched origin/main=692735e91d6a9ab07ddd3686feb60a56b270728a.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: The dev service is an rdm build, so startup skipped model seeding and kept the old MiniMax chat model. Later starts need the saved env file applied even without an admin key command.
Evidence: seed and rdm launcher tests passed. The live dev database now uses the four-model chat route and keeps the original MiniMax endpoint as reserve.
Impact: Restart of a service with envfile set rewrites system model settings. HTTP seeding is unchanged for builds that can mint an admin key. MiniMax is stored separately because fallbacks cannot switch gateways.
Refs: .task/feature-four-model-vision.md
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: deploy can compile selected plugins into one service binary.
Evidence: Actions run 3288 jobs 5053 test, 5054 test-pg, and 5055 svc-realcombo success on 77f7975cac. Head contains origin/main 63e70d8.
Impact: default remains no plugins; --with selects the catalog. Lifecycle suite stays NOT_RUN.
Refs: #558
Why: branches merged into main had left their task cards open; backfill status and archive to prevent bookkeeping debt.
Evidence: feature-pptx-gate-reference-integrity.md(PR#559) feature-workbench-visual-redesign.md(PR#556) — merges verified via git log --merges plus Gitea API squash lookup against origin/main.
Impact: .task/*.md moved to .task/archive/ with status backfilled; no code or submodule change.
Refs: task.sh close (card-autoclose-r1)
Why: include the current integration baseline before verification.
Evidence: fetched origin/main. The multi-instance account bootstrap conflict keeps main's identity ensure and activation flow.
Impact: integrate upstream changes without rewriting task history. Backend gitlink stays 80bb123.
Refs: task-sync
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Dev CI must build the backend head that contains current main and the child-event assertion fix.
Evidence: Agent-System-Backend feature/svc-plugin-host-options is 80bb123.
Impact: modules/backend gitlink only.
Refs: Agent-System-Backend 80bb123c874cd9eb0e25f12ccd41e888f727a7e1
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: include the current integration baseline before verification.
Evidence: fetched origin/main=59cbf1109634b66c88f5b1702f330cd738aa75ef.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
Why: The multi-instance acceptance test still created alice with a literal
password; backend U3 replaced local password-bearing creation with trusted
pending-account activation, so the first pin bump past U3 failed test-pg
with RECOVERY_CONTRACT_REPLACED. The seam was left pending by the U3 task
and this PR is the first to cross it.
Evidence: Full TestTwoInstancesOverSharedPG passes against a disposable
PostgreSQL 16 container (identity ensure -> bootstrap-admin -> pending
create with activation key -> key redemption -> grant-setup password ->
api key -> all cross-instance scenarios, 15s).
Impact: Test-only bootstrap rewrite (deployment-level identity ensured once
per pair, admin bootstrapped first, alice activated through the key
redemption and grant-setup password flow); no product code changes.
Refs: Dev task feature/pptx-gate-reference-integrity; backend U3 #198
Why: the current heads are not green, so the merge chain stops before any repository is merged.
Evidence: SELF query of the open pull requests and their current-head Actions runs.
Impact: task card only. This commit stays local so the open Dev pull request head does not move.
Refs: Agent-System-Dev pulls/558
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: the backend branch moved forward to include the OOXML gate fix, and this checkout should name that commit.
Evidence: Agent-System-Backend feature/svc-plugin-host-options is bffd37c.
Impact: modules/backend gitlink only.
Refs: Agent-System-Backend bffd37c42f809332499d3227ba1377c8d40c9459
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: The review candidate needs reproducible evidence bound to the exact Dev and Frontend commits.
Evidence: SELF receipts record 272 unit tests, the production build, and 16 browser scenarios on Dev 49ef24b.
Impact: Updates only the task record; product behavior is unchanged.
Refs: feature/workbench-visual-redesign
Why: bootstrap-admin now refuses a database that has no agent_system_identity row, and serve does not insert one.
Evidence: Dev CI run 3266 job 5008 failed with system identity missing on 7eef723.
Impact: test setup only.
Refs: Agent-System-Dev pulls/558
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: the synced backend rejects password-bearing user create, so the shared-PG acceptance could not open its first account.
Evidence: Backend CI run 3261 succeeded on 8c60ed0; Dev test-pg run 3263 failed in TestTwoInstancesOverSharedPG before this change.
Impact: test setup only. Alice remains a non-admin account and still receives an API key after setting a password.
Refs: Agent-System-Dev pulls/558
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: The review candidate needs reproducible evidence bound to the exact Dev and Frontend commits.
Evidence: SELF receipts record 272 unit tests, the production build, and 16 browser scenarios on Dev 49ef24b.
Impact: Updates only the task record; product behavior is unchanged.
Refs: feature/workbench-visual-redesign
Why: include the current integration baseline before verification.
Evidence: fetched origin/main=cc40e10f3b4b6e41cb3a7aa0826e08dd3b8b7270.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
Why: The review candidate needs reproducible evidence bound to the exact Dev and Frontend commits.
Evidence: SELF receipts record 272 unit tests, the production build, and 16 browser scenarios on Dev 49ef24b.
Impact: Updates only the task record; product behavior is unchanged.
Refs: feature/workbench-visual-redesign
weichenandCursor