Why: the U4 branch is 89 commits behind main, which now carries the shared
document renderer, two-trees project work, the SQLite instance guard and
exec env layers (PR #209 merged as c76d8bd); it conflicted in storage
migrations, host capabilities and the extension surface. The user authorized
syncing U4 after the renderer line on 2026-10-18.
Evidence: content is the agent-exec transport 3b64eb5 (cursor worker, unit
backend-iam-u4-onto-main-c76d8bd-r1, run on a checkout whose tree equals main,
receipt /tmp/cleanup-20261018/render/u4-backend-onto-run.json, SELF). Host
check: main migrations 1-48 unchanged (migrate.go has no deleted lines versus
main), U4 migrations moved from 35-36 to 49-50 with identical SQL,
surface.txt is the sorted union, deps stay at main's pins.
Impact: databases that ran unmerged U4 builds with the old numbering must be
recreated. SurfaceVersion becomes 2.25.0-experimental. Main's admin password
account creation now requires the system.users.create action like the other
U4 account writes.
Refs: Backend main c76d8bd, PR #209
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: the previous commit archived the card without its closure line.
Evidence: card text only.
Impact: none outside .task.
Refs: PR #209
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: PR #209 delivers this W9 increment; the card must not stay open on main.
Evidence: Backend HEAD 84a6f6f CI vet-and-test and governance succeeded (GATE).
Impact: card archived; records the 37-48 migration renumbering and the user-owned v0.6.0 tag.
Refs: PR #209
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: PR #209 is behind main by 14 commits (admin passwords, managed
interpreter path, U8 SQLite instance guard, EXT-20 regulations) and conflicted
in app wiring, host capabilities, migrations and the extension surface; the
user authorized merging the renderer line on 2026-10-18.
Evidence: content is the agent-exec transport 87a99ece (cursor worker, unit
backend-task10-sync-main-56a7f9f-r2, receipt
/tmp/cleanup-20261018/render/backend-sync-run2.json, SELF): gofmt, go build,
go vet and go test ./... (43 packages) passed. Host check: main migrations
1-36 unchanged, feature migrations moved from 36-47 to 37-48 in order with
identical SQL, surface.txt is the sorted union.
Impact: main's migration 36 instance_guard stays authoritative; databases that
ran unmerged feature builds with the old numbering must be recreated.
SurfaceVersion becomes 2.24.0-experimental.
Refs: PR #209, Backend main 56a7f9f
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Agent-Document-Renderer main publishes company.default@1.1.0 with a
declarative TOC (spec 04, acceptance 10 §4 §7) and keeps 1.0.0 loadable by
version. The fixed defaultProfileVersion "1.0.0" no longer matched
LoadCompanyDefaultContext (1.1.0), so every delivery failed with a profile
mismatch. New documents must take the latest published version, while
documents that already bind a published version keep rendering with exactly
that version, because published rules are immutable and old versions remain
retrievable (spec 08).
Evidence: go vet ./... ok; go test -count=1 ./... ok (all packages);
go run ./internal/docgen/internal/genrendereridentity -check ok;
scripts/agent-exec validate --repo . ok; scripts/dm governance check --repo .
ok; gofmt -l clean; git diff --cached --check clean. New tests cover the
default binding (1.1.0, TOC present), pinned 1.0.0 (no TOC) and 1.1.0 for
both a request pin and a persisted-Markdown replay, and refusal of unknown
versions and other profile IDs before any profile load.
Impact: docgen resolves the profile version from the persisted Markdown
binding or the request, defaults new documents to
profiles.LatestCompanyDefaultVersion, and loads it through
LoadCompanyDefaultVersionContext. The legacy migration comparison records the
declared TOC as an intentional corporate-layout difference
(RDR-T10-CMP-002) and keeps it out of the body visible-text dimension.
RendererIdentity is regenerated for 3f259990b5d4.
Refs: unified renderer plan Task 13; renderer PR #7
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Bring feature/unified-document-renderer-task10 up to date with main
(PR #208 governance kit adoption: deps/development-guidelines and
deps/development-manager gitlinks, AGENT-EXECUTION.md, scripts, CI workflow).
Evidence: .gitmodules conflict resolved by keeping all three new submodule
entries (agent-document-renderer from this branch; development-guidelines and
development-manager from main). go vet ./... ok; go test -count=1 ./... ok
(42 packages); genrendereridentity -check ok; scripts/agent-exec validate ok;
scripts/dm governance check ok.
Impact: No code changes beyond main's governance files; renderer gitlink
stays at 1095077.
Refs: origin/main cac60c0
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: the adoption PR is reviewed and its candidate HEAD passed CI; the card closure is folded into the PR so no separate close PR is needed.
Evidence: CI GATE on 19e86b5150e019e45f86ee9a97c6dfa153654d06; independent Review REPLAY (codex exec --ephemeral（fresh_session=true，session id not-observable）VERDICT: APPROVE，无阻塞项).
Impact: task card only (moved to .task/archive/).
Refs: https://git-home.u1680832.nyat.app:30319/weichen/Agent-System-Backend/pulls/208
Why: Phase 6 requires one positive executor run and the negative cases (missing config, bad receipt, out-of-scope path, unknown side effect) before adoption becomes active.
Evidence: .task/evidence/governance-adoption/ (SELF): governance check, agent-exec validate and adoption cases; this commit passed the generated pre-commit hook (admit).
Impact: task card and evidence logs only.
Refs: .task/feature-governance-adoption.md
Why: development-governance-v2 Phase 6 adopts the shared governance kit in every formal repository; the executor needs a valid AGENT-EXECUTION.md before any implementation unit can run.
Evidence: generated by dm-server governance init at Development-Manager bd3a7de57a3a9e1a4495b94c5fa5c08d71e7be8c; dm-server agent-exec validate passes.
Impact: configuration and task card only; no code or CI change yet.
Refs: Agent-System-Dev docs/superpowers/plans/2026-10-16-development-governance-v2-master.md
Why: agent-document-renderer 1095077 (origin/main 10950774) adds node-located
content diagnostics (required_content_missing, intentionally_empty,
template_placeholder_left, mapping_fallback, signature_data_missing,
mapping_unresolved, mapping_ambiguous), a mapping Required flag, the
intentionally-empty model property, refusal of approved output with missing
required content or trusted signature data, expansion of empty
w:sdtContent so content lands inside content controls, and nested mapped
content controls rendered inside parent slots. Backend pinned 31fbe54.
Evidence: go generate ./internal/docgen and go run
./internal/docgen/internal/genrendereridentity -check ok; go vet ./...
exit 0; gofmt -l on changed Go files empty; go test -count=1 ./... exit 0
(42 packages ok, including renderer integration tests in internal/docgen);
renderer submodule go test -count=1 ./... ok at 10950774. New
TestRendererNodeDiagnosticsPassThroughUnchanged asserts renderer codes,
severities, NodeID and RulePath pass unchanged through the manifest and
DiagnosticError. git diff --cached --check clean.
Impact: Renderer identity and the published pin move to 10950774736b64b5d8e544ce03d45ae5a6ceca7c.
Backend only renders ordinary workspace working/preview documents; approved and
review modes still route to RDM before the renderer runs, so the approved
refusal paths are not reachable from Backend. Typical Backend documents produce
no new diagnostics and their DOCX output is unchanged in the test suite. Node
diagnostics on the failure path are still returned only as DiagnosticError
values and are not added to the operation receipt, which carries error text
only; that contract is unchanged.
Refs: unified renderer plan Tasks 10/13
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: agent-document-renderer 31fbe54 publishes company.default@1.0.0 with
review status approved and publicationAllowed true (owner decision
2026-10-16: A4, ~2.5cm margins, SimSun + Times New Roman 12pt body, SimHei
headings, no logo). Backend still hard-coded Candidate=true and every prompt,
skill and tool text called the default a pending-human-visual-review
candidate.
Evidence: go vet ./... exit 0; gofmt -l on changed packages empty;
go test -count=1 ./internal/docgen/... ./internal/prompts/...
./internal/skills/... ./internal/tools/... ok; go test -count=1 ./... exit 0
(42 packages ok); go build ./... and CGO_ENABLED=0 go build ./cmd/agent-system
ok; git diff --cached --check clean. New tests failed before the change.
Impact: bump deps/agent-document-renderer 0d1c512 -> 31fbe54 and regenerate
the renderer identity. The manifest profile receipt now reports the
renderer's own review status and derives Candidate from the snapshot state
(false for the published default). Prompts (base.deliverables v12,
tool.render_document v2, task.report v4, task.doc-minutes v4), docx-author
and house-style skills, the render_document tool description and comparison
exception details describe the default as published (approved by owner
2026-10-16) while still limiting it to ordinary uncontrolled workspace
working/preview Word. Routing is unchanged: project, controlled,
special-template, review and approved outputs still go to RDM, with no
Backend default fallback and no legacy Word builder fallback.
Refs: unified renderer plan Tasks 7/10
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: The registration gate classified Word packages from a hard-coded ZIP member path, allowing alternate main-part paths to bypass the gate and rejecting unrelated ZIP archives.
Evidence: Focused OPC registration tests; go test -count=1 ./internal/tools; go test -count=1 ./...; go test -race -count=1 ./internal/tools ./internal/runmgr; go vet ./...; go build ./...; git diff --check.
Impact: Resolve package-level officeDocument targets safely, classify all conventional Word main content types through OPC defaults or overrides, and fail closed on malformed, ambiguous, external, or traversing metadata while preserving non-Word artifacts.
Refs: Task 10; /tmp/task10-backend-final-closure-last.md
Why: Direct artifact registration trusted filename extensions, the shared renderer parity test used a reduced fixture, and migration comparison omitted canonical list/task/rule/alignment properties plus their Word presentation.
Evidence: go test -count=1 ./...; go test -race ./internal/docgen ./internal/tools ./internal/runmgr; go vet ./...; go build ./...; PYTHONPATH=. python3 -m unittest tests.test_housestyle; git diff --check.
Impact: Reject renamed WordprocessingML packages while preserving non-Word artifacts, pin the reviewed renderer semantics, and report canonical and DOCX presentation regressions as accidental differences.
Refs: Task 10; /tmp/task10-backend-final-review-last.md
Why:
- Independent review found trusted-boundary, semantic parity, replay-budget,
comparison, persistence, cancellation, provenance, and routing gaps.
- Storage migration must include the new durable delivery table.
Evidence:
- go test -count=1 ./...
- go test -race ./internal/docgen ./internal/tools ./internal/runmgr
- go test ./... (deps/agent-document-renderer)
- go vet ./...
- go build ./...
- CGO_ENABLED=0 go build -trimpath ./cmd/agent-system
- Python house-style unittest suite and git diff --check
Impact:
- Word delivery now uses host-owned freezes, Markdown-first resumable persistence,
closed registration boundaries, semantic shared rendering, and real comparison.
- Controlled/project documents still fail closed to RDM and non-Word flows remain
unchanged.
Refs:
- Unified Document Renderer implementation plan Task 10
- /tmp/task10-backend-review-last.md
- Renderer d5cd8c126045a27d5a5cc5da23f64d3c737e116a
Why: The Task 10 corrective pass requires the shared Renderer semantic Markdown model and synchronous cooperative cancellation APIs.
Evidence: Renderer go test ./... and Backend go test ./internal/docgen -count=1 passed at d5cd8c126045a27d5a5cc5da23f64d3c737e116a.
Impact: Advances only the immutable agent-document-renderer gitlink; Backend repair behavior is committed separately.
Refs: Unified Document Renderer Task 10 review repair; /tmp/task10-backend-review-last.md
Why: Agent-facing prompts and house-style recipes must exercise the shared renderer instead of leaving the old Python Word builder as the ordinary production route.
Evidence: Prompt routing tests, house-style Python tests, affected-package race checks, full Go tests, vet, builds, and static binary verification passed locally.
Impact: Word guidance now calls render_document, labels the pending organization-default candidate, routes controlled/project work to RDM, and gates housestyle.build behind AGENT_SYSTEM_LEGACY_WORD_COMPARISON=1.
Refs: Unified Document Renderer Task 10; Renderer d449b9f779f2f968f8ac556b89940a6c9aaf3fdc
Why: Task 10 requires Backend to consume the reviewed canonical renderer and its embedded organization-default candidate without copying assets.
Evidence: Fresh GOMODCACHE resolution loaded agent-document-renderer/profiles and passed its package tests at d449b9f779f2f968f8ac556b89940a6c9aaf3fdc.
Impact: Adds one immutable Renderer gitlink plus the local Go module require/replace; no production path changes yet.
Refs: Unified Document Renderer Task 10; Renderer d449b9f779f2f968f8ac556b89940a6c9aaf3fdc
Why: Integrate Backend origin/main 7aa6243 while retaining the product's mandatory trusted project-context startup contract.
Evidence: go test ./pkg/extension/... ./internal/app -count=1; go vet ./...; go test ./... -count=1; go build ./...; CGO_ENABLED=0 go build ./cmd/agent-system.
Impact: Keep the appOptions helper and exec-environment fallbacks while forwarding RequireProjectContext without loss.
Refs: 92f9286e6c4004074f014902f8b04554287ddfb2; 7aa624370a
Why: Integrate the exact Dev main pin while preserving trusted project context, vision routing, work-result context, and distinct migration identities.
Evidence: go vet ./...; go test ./... -count=1; go build ./...; CGO_ENABLED=0 go build ./cmd/agent-system; isolated PostgreSQL migration history tests.
Impact: Keep exec_env_layers at schema v35, move the feature migrations to v36-v46 in order, and fail closed on the old unpublished numbering.
Refs: 14e780c6fa32fdd4742b6232150c5efcb0d766ff; ab12fea474
Why: The real local-mode product has authoritative project access without an enabled account-model row.
Evidence: The app suite passes; non-local missing or disabled accounts still fail closed.
Impact: Auth-off local acceptance can record a nonempty authorization epoch while authenticated deployments retain account-version binding.
Refs: TT-N03
Why: A generic conflict response hid whether source intake, authorization, or risk analysis blocked a formal command.
Evidence: The server suite passes with bounded domain reasons included in the conflict response.
Impact: Clients can present the actionable gate reason without exposing source contents.
Refs: TT-N03 TT-R01
Why: Direct HTTP and surface command dispatch could mutate RDM project facts without an analyzed committed-message manifest.
Evidence: Command-bus, server, and app suites pass with a classification regression for formal writes and read/proof exceptions.
Impact: Formal HTTP writes require a conversation anchor and committed proceed decision; surface writes fail closed until they can carry the same proof.
Refs: TT-N03 TT-R01
Why: Direct search, exec, fetch, and MCP providers could return project-driving results without entering the trusted source observation gate.
Evidence: App and run-manager suites pass with project runs confined to the audited runCtx tool path.
Impact: Non-project native runs retain configured tools; project runs fail closed until native result hooks can prove durable observation.
Refs: TT-N03 TT-R01
Why: Accepted retrieval and tool sources could become stale, and a pending HITL answer could release a changed source set.
Evidence: App and run-manager suites cover per-preflight resource verification, authorization epochs, and a second exact-source preflight after HITL.
Impact: Project work fails closed when authorization, source bytes, run binding, or the source watermark changes.
Refs: TT-N03 TT-R01
Why: The W4b verifier branch must include the current compiled-plugin host option surface before review.
Evidence: Focused extension, app, server, and command-bus tests pass after preserving both public API additions.
Impact: The public surface advances to 2.23.0-experimental and exposes both human-decision and compiled-plugin capabilities.
Refs: W4b human decision verifier
Why: An interrupted import cleanup removed the cmdbus core import while two human-decision files used an explicit alias.
Evidence: go test ./internal/cmdbus ./internal/server passes.
Impact: The backend worktree is buildable and consistently names the core package without behavior changes.
Refs: W4b human decision verifier
Why: The new confirmation table is additive, and the compatibility ledger must name that fact or the store suite rejects the schema.
Evidence: SELF go test ./internal/store/ -run TestMigrationCompatRegistered passed after the registration.
Impact: Old readers keep their current floor. They do not need to understand human decision rows.
Refs: W4b human confirmation design r1
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: A model-visible flag cannot prove that a person confirmed one proposal revision and payload hash.
Evidence: go test ./pkg/extension ./internal/app ./internal/server -run 'HumanDecision|SurfaceSnapshot|ResolveHuman|NaturalLanguage' passed locally.
Impact: SurfaceVersion 2.22.0-experimental adds HumanDecisionVerifier. Existing Host methods stay unchanged, and the base does not consume the proof.
Refs: W4b human confirmation design r1
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Independent review found delegated account actions, stale resource policy, and role migration cutover gaps in the initial U4 implementation.
Evidence: Backend full SQLite suite passed; isolated PostgreSQL role and migration suite passed; independent final review found no remaining findings.
Impact: Enforces transactional system grants, current role snapshots, staged role import with v36 migration, and fail-closed activation.
Refs: feature/iam-u4-role-authority
Why: Non-root managers need the system-management entry and controls to follow their explicit grants.
Evidence: The full Backend test suite passed and /api/me action projection is covered by the grant-isolation test.
Impact: /api/me now returns only the current actor system actions evaluated by the same authorizer.
Refs: Focused role and permission system task 7
Why: Existing RDM role commands must reach the Backend authority during UI and migration cutover without dual-writing RDM role tables.
Evidence: The full Backend suite passed, including actor-bound role administration, owner compatibility, provider, and public-surface tests.
Impact: Adds temporary trusted Host administration methods and completes the project-action catalog used by RDM.
Refs: Focused role and permission system task 5
Why: RDM needs a trusted, non-secret view of Backend-owned role policy without importing Backend internals or accepting a caller-supplied actor.
Evidence: The full Backend test suite passed, including public-surface and provider contract tests.
Impact: Adds optional extension surface 2.14 with actor-bound role snapshots and action decisions.
Refs: Focused role and permission system task 4
Why: System, role, and project entry permissions need explicit actions instead of a single administrator flag or unrestricted project creation.
Evidence: The full Backend test suite passed, including role APIs, action isolation, project.create denial/allow, storage migration coverage, and account write fences.
Impact: Adds versioned role and grant endpoints, maps account actions to explicit permissions, and requires project.create authority.
Refs: Focused role and permission system task 3
Why: Project messages, retrieved evidence, tool results, and final replies must not affect model work or reach users before the domain risk decision is complete.
Evidence: RED tests proved the runner, final reply, and message deltas advanced before the gate; focused tests, go test -p 1 ./..., go build ./..., and isolated PostgreSQL runtime tests now pass.
Impact: Wire the optional RDM observer into project run start, retrieval, tool-result, HITL clarification, and final reply buffering; hosts without the optional provider keep existing behavior.
Refs: TT-N03, TT-R01; approved work-observation design and stage-five plan
Why: Role assignment and permission changes must preserve one active position, version conflicts, idempotency, and the final policy manager.
Evidence: The full Store suite passed on SQLite and focused role/policy tests passed on isolated PostgreSQL.
Impact: Adds transactional role mutations and seeds explicit system grants for existing administrators.
Refs: Focused role and permission system task 2
Why: Backend needs a single trusted catalog and durable role authority before RDM can stop owning position data.
Evidence: Authz catalog tests and the full store suite passed on SQLite; migration rollback and one-active-role constraints are covered.
Impact: Adds schema v35 and stable role/binding types without changing existing authorization callers.
Refs: Focused role and permission system task 1
Why: Project retrieval hits, tool results, and saved visible analysis must be individually traceable and fully analyzed before later project actions consume them.
Evidence: RED tests first showed missing run-source capture, tool-result capture, and run-wide preflight; go test -p 1 ./... and go build ./... passed; isolated PostgreSQL source, migration, audit, and context tests passed.
Impact: Persist actual run-bound sources with current-status checks and exact manifests behind the disabled X3 assembly gate; no live product path, migration, deployment, or release is enabled.
Refs: TT-N03, TT-R01; approved X3 stage-five plan r1
Why: A durable observation receipt must not be mistaken for completed risk analysis or permission to execute a project action.
Evidence: Exact-source preflight tests cover pending versus proceed, partial attachment, post-ACK source tampering, current authorization, and receipt idempotency; go test -p 1 ./... and go build ./... passed.
Impact: Add a fail-closed host preflight helper and per-message manifest read, still unwired to the public product path pending recovery and four-route coverage.
Refs: TT-N03, TT-R01; approved X3 design r1
Why: Project user messages and final assistant replies need exact, replayable source parts before a domain observer can safely analyze them.
Evidence: go test -p 1 ./... passed; isolated PostgreSQL outbox, audit, migration, and work-context tests passed against a temporary cluster; request dedup, fenced settlement rollback, authorization recheck, and receipt idempotency tests passed.
Impact: Add migration 44 and a private host outbox for committed message parts; actual observer delivery and action preflight are not yet wired into the product path.
Refs: TT-N03, TT-R01; approved X3 stage-five plan r1
Why: X3 must share one Backend surface with the X1 context and X2 audit candidates while current main now owns account migrations 32-34.
Evidence: go test -p 1 ./... and go build ./... passed on the merged tree; isolated PostgreSQL project audit, migration, and work-context tests passed against a temporary instance. Default parallel go test ./... had only the unrelated sandbox orphan timing test fail; its isolated ten-run replay passed.
Impact: Preserve account migrations 32-34, place X1/X2 migrations at 35-43, and retain account-effective write guards alongside project audit/context behavior. X3 observer remains a contract only.
Refs: TT-C01-TT-C04, TT-A01-TT-A04, TT-N03, TT-R01; combination probe d9b00c2
Why: Real project messages and retrieved material need a scoped observer and preflight decision contract before they can become governed project candidates.
Evidence: Registration denial test was RED before the implementation; go test ./... passed on Backend origin/main 2716b9e.
Impact: Add public X3 types and trusted single-provider registration; the production observer slot is deliberately not wired yet, so no project observation behavior is claimed.
Refs: TT-N03, TT-R01; approved work observation stage-five plan r1
Why: An already-open Store could otherwise execute a numbered down migration after its ledger name changed.
Evidence: The targeted rollback test failed before the guard and passed after it; focused SQLite and isolated PostgreSQL store/migrator checks passed.
Impact: MigrationStatus and MigrateDown fail closed on version-name mismatch before changing schema.
Refs: feature/rdm-two-trees-latest-templates Backend combination probe
Why: Combined W6 and W9 storage validation must cover actual work task and step-source rows, not only empty tables.
Evidence: Isolated PostgreSQL migration test copies and reads both rows with their trusted scope intact.
Impact: Adds integration coverage for the combined v31-v40 migration; no production behavior change.
Refs: feature/rdm-two-trees-latest-templates Backend combination probe
Why: Parallel unpublished candidates reused numeric versions, so version-only checks could reinterpret a different schema after combination.
Evidence: SQLite ledger and source-copy mismatch tests failed before the guard and pass now; PostgreSQL store/migrator and full Go suites pass.
Impact: Fail closed on version-name mismatch before opening a migration target or serving a database.
Refs: feature/rdm-two-trees-latest-templates Backend combination probe
Why: Both approved Backend slices share extension surfaces and migration numbers, and current main added account audit v31.
Evidence: The isolated probe resolves additive interfaces and v31-v40 ordering; focused and full Go tests passed without PostgreSQL.
Impact: Produces a local integration candidate for isolated PostgreSQL and race verification only; shipping branches remain unchanged.
Refs: feature/rdm-two-trees-latest-templates W6 W9
Why: Validate W6 against account-audit migration v31 on the current authoritative base.
Evidence: Focused store/runmgr/migrator and full Go tests passed in the isolated probe after resolving migration collisions.
Impact: Renumbers unpublished W6 migrations to v32-v36; this probe commit is not a shipping branch or live migration.
Refs: feature/rdm-two-trees-latest-templates W6 integration probe
Why: Explicit low IDs bypassed the implicit sequence allocator and could commit after a higher audit page cutoff.
Evidence: Isolated PostgreSQL RED receipt for cross-page omission; local GREEN for explicit and default writes, trigger drift, isolation, mutation denial, offline copy rollback, retry, and sequence non-rewind. Fixed-HEAD receipts follow.
Impact: Add v34 all-writer and append-only guards; reject unsafe strong reads; make command audit storage migration exclusive and atomic with guard restoration and a forward-only sequence.
Refs: Approved W9 all-writers cutoff adjustment r1; W9 Backend task card.
Why: A crash after dropping candidate raw prompt columns but before SQLite compaction left physical prompt bytes without a visible reason to retry, and storage backup could accept that file.
Evidence: RED post-commit interruption and pending-backup tests; full Go suite, targeted race, isolated PostgreSQL package tests, vet, and build passed locally before commit.
Impact: Commit a durable pending marker with v35, retry compaction and WAL truncation until completion, conservatively re-scrub earlier v35 candidates, and reject backup or migration while cleanup is pending or unverified.
Refs: Approved RDM stage6 W6 source-recheck design adjustment r1; W6 independent review crash-window finding.
Why: PostgreSQL sequence IDs could commit out of order and make fixed project audit cursors permanently skip a late row.
Evidence: SELF PG cross-page RED, CACHE 10 counterexample, configuration drift RED/GREEN, and pre-commit Go, PG migration, and race receipts are recorded in the W9 task card. Fixed-HEAD checks follow this commit.
Impact: Guard implicit command audit ID allocation with a transaction advisory lock; reject unsafe sequence/default drift; preserve explicit-ID storage migration; report committed-prefix only for verified project pages.
Refs: Approved W9 PostgreSQL cutoff adjustment r1; W9 Backend task card.
Why: Goal-sensitive workflow verification persisted model prompts and prior output text in trusted receipt rows, while cross-step model inputs could reuse stale machine output.
Evidence: RED receipt privacy test; full Go suite, targeted race, isolated PostgreSQL store/runmgr/storage-migrate tests, vet, and static build passed locally before commit. A known unrelated TempDir cleanup flake failed once and passed on targeted and serial full replay.
Impact: Add optional same-provider source recheck, verify every included item and evidence body before reuse, gate cross-step outputs, reject unsettled attempts, remove raw scope columns, scrub isolated candidate v34 SQLite/PG schemas, and reject unsafe backups.
Refs: Approved RDM stage6 W6 source-recheck design adjustment r1; W6 independent review.
Why: A goal-sensitive project provider can return different authorized sources for a prompt step and its judge, causing a valid workflow to fail when the judge re-resolves with an empty goal.
Evidence: RED goal-sensitive step test; full compiled workflow positive and withdrawal negative tests; go test ./... -count=1 and targeted race passed locally before commit.
Impact: Persist each trusted model attempt selector outside the public manifest, enforce receipt state continuity, and reauthorize every contributing model turn with its original selector before judging.
Refs: RDM two-trees stage5 W6; W6 independent review P2.
Why: Compiled workflow critics could resend a prior prompt output after document duty or publication changed across separate model guards.
Evidence: RED cross-guard revocation and retry tests; go test ./... -count=1 and targeted go test -race passed locally before commit.
Impact: Bind materialized prompt bytes to the engine step attempt and responded model receipt; recheck current project sources and epochs before judging, failing closed for unproven or ambiguous output.
Refs: RDM two-trees stage5 W6; stage6 X1 public extension request; W6 independent review P1.
Why: Revoked project material could remain in later model turns, and plugin Audit could impersonate a model-input manifest.
Evidence: SELF red tests reproduced both paths; updated project revocation, trusted receipt, migration, and entrypoint tests pass locally with go test ./... -count=1 and targeted race checks.
Impact: Rebuild direct model requests after scope changes, sanitize or block unproven generic history, omit legacy project injections, and persist manifests in a dedicated state-checked table.
Refs: TT-C02 TT-C03 TT-C04; stage5 W6; review P1 stale context and manifest provenance
Why: Clock-derived IDs could collide and string-scoped hosts could impersonate a registered audit publisher.
Evidence: Fixed-clock concurrent, ambiguous private offset, forged scoped-host, and registered publisher RED-to-GREEN tests; go test ./... -count=1 and targeted race test passed locally as SELF.
Impact: Generate 192-bit host IDs, reject ambiguous explicit offsets, and issue a publication-only capability at verifier registration.
Refs: Stage 5 W9; approved Stage 6 audit-link design adjustment r1; review of a8fa2fe
Why: A domain outbox could not correlate its committed event with the host receipt because the bus generated the operation ID after the handler returned.
Evidence: RED-to-GREEN bus and compiled-handler tests; local go test ./... -count=1 and isolated PostgreSQL operation/link checks passed as SELF.
Impact: Expose a host-generated operation ID to handlers and resolve a unique committed audit offset after dispatch; missing receipts remain retryable and ambiguous rows fail closed.
Refs: Stage 5 W9; approved Stage 6 audit-link design adjustment r1
Why: Project work currently reaches model calls without current authorized project facts or a persisted input receipt.
Evidence: SELF go test ./... -count=1, targeted go test -race for W6 paths, go vet ./..., and CGO_ENABLED=0 go build ./cmd/agent-system passed on this change set.
Impact: Add optional X1 interfaces, trusted persisted work-task binding, request-scoped context injection, evidence reauthorization, and attempted/responded/unknown manifests. Existing non-project behavior stays available.
Refs: TT-C01 TT-C02 TT-C03 TT-C04; stage5 W6; Dev stage6 X1 public extension request r1
Why: Owner-only command audit cannot prove shared business commits, and button dispatch omitted its authenticated actor.
Evidence: Local go test ./... passed; isolated PostgreSQL cutoff, link, and SQLite-to-PostgreSQL migration receipts are recorded as SELF in the task card.
Impact: Add an optional domain-verified metadata link, exact authorized cross-member read, idempotent storage, actor propagation, and honest PostgreSQL cutoff coverage. Broad audit pagination and other sources remain pending.
Refs: Stage 5 W9; approved Stage 6 audit-link design adjustment r1
Why: Project audit cannot use owner-wide logs or current conversation project state to establish trusted attribution.
Evidence: RED cases were observed locally; go test ./pkg/extension ./internal/store ./internal/app -count=1 passed on this tree (SELF).
Impact: Add optional X2 read and audit.read callback, immutable project/name snapshots, safe owner-only projection, and explicit unsupported coverage. Cross-member publication remains pending.
Refs: Stage 5 W9; Stage 6 X2 public extension request
Why: svc.sh delivers database credentials via environment per G1-7
(command lines are visible to every process on the host), but identity
ensure/verify only accepted --dsn on the command line for pg-bound
instances.
Evidence: go test ./pkg/identity/... ok; build ok.
Impact: --dsn remains highest priority, then AGENT_SYSTEM_DSN, then
embedded SQLite; no storage changes.
Refs: Dev U3 (real-combination CI gates); G1-7 credential discipline.
Why: runIdentity lived in cmd/agent-system package main, so downstream
combination binaries (Agent-System-RDM) cannot reuse it and their unknown
subcommands fall into the serve path; the 2026-10-12 rdm1/rdmcand batch
update failures traced back to that capability gap.
Evidence: go test ./pkg/identity/... ok; CGO_ENABLED=0 go build ok;
behavioral probes (usage rc=2, capabilities v1, ensure/verify roundtrip)
recorded in the PR.
Impact: identity semantics unchanged for the base binary; adds
identity capabilities (v1) and usage-on-unknown contract shared by all
products; no storage changes.
Refs: Dev PR #474 (svc.sh candidate gate); user root-cause report
2026-10-12 U1 (shared identity interface).
Why: pluginRegistry never overrode RegisterResourceAuthorizer, so
in-proc extensions hit the declarative-branch rejection; register()
aborted wholesale and every plugin command vanished from the bus. The
RDM binary ships the EXT-8 member adapter since step-4, meaning the
served RDM process answers 'unknown or disabled command' for every
rdm.* call — caught by the MS1 re-acceptance smoke replay (fb/report/
confirm/lifecycle smokes all failed on rdm.baseline).
Evidence: TestInprocPluginRegistersAuthorizerAndCommands — red on the
parent commit (registration aborts, dispatch rejected, nil-slot denies),
green with the fix (command committed; the shared authz slot answers
alice-allow/bob-deny through projectAuthz). Full go test ./... green;
RDM MS0 acceptance re-run follows on the pin bump.
Impact: pluginRegistry.RegisterResourceAuthorizer fills the shared
hostcaps.authz slot (first registrant wins, mirroring the composition
root); compiled/declarative branch unchanged.
Refs: RDM G1-8 MS1 re-acceptance; EXT-8 (d89e748 seam); RDM #113.
Why: Dev umbrella pin consistency requires backend's agent-squad gitlink
to match Dev's; squad main moved to 8bbf850 (nested agent-native pin at
3b92cb63, same as backend's own native pin after #115), so backend must
carry it before Dev can consume 8157e68+.
Evidence: gitlink-only change; Squad #49 merge 8bbf850 records
deps/agent-native=3b92cb63 (verified via ls-tree).
Impact: nested gitlink only; no backend source change.
Refs: Squad PR #49; Dev check-pins cascade; Backend #115/#116.
Why: the EXT-14 wiring made runmgr import cmdbus for the CommandSource
interface signatures, which adds a new edge backend/runmgr ->
backend/cmdbus outside the architecture baseline ratchet; the Dev
umbrella lint (lint-architecture) rejects the gitlink bump because of
it. The dependency direction was always meant to be inverted: runmgr
defines the port, the composition root adapts the bus.
Evidence: go build/vet clean; go test ./... 32/32 packages green;
grep confirms no internal/cmdbus import left in non-test runmgr files;
Dev-side lint re-run scheduled on the gitlink bump PR.
Impact: runmgr.CommandTool/CommandRequest/CommandResponse are local
view types; app.busAdapter now converts cmdbus types at the boundary
(Entrance still fixed to agent by runmgr). No behavior change; the
runmgr tests speak the view types instead of cmdbus ones.
Refs: Dev umbrella make test failure 'architecture: new dependency
edges backend/runmgr backend/cmdbus'; proposal 2026-10-09-ext-g1-b1-b3
(EXT-14); Backend #114.
Why: the adjudicated preview plan needs server-side OG fetching for link
cards, and the proposal's reserved-capability items call for the fs tool
family to exist behind a default-off gate (implemented, not offered).
Evidence: GET /api/url-preview fetches and parses og:title/description/
site_name/image/favicon with a 5s timeout, 256KB body cap, text/html
content check, TTL cache, and SSRF guards enforced at dial time (loopback
/private/link-local/multicast/reserved/CGNAT rejected per hop, anti DNS-
rebind; AS_URL_PREVIEW_ALLOW_PRIVATE=1 exists for local testing only,
default fail-closed; errors return sanitized cards, never internal
addresses). fs_read_file/fs_write_file/fs_list_dir register as Hidden
tools sandboxed under <root>/<owner>/<conversationId>: traversal,
absolute paths, backslashes, and symlink escape rejected; 1MB/2MB size
caps; writes register an artifact so workspace files stay visible in the
panel. ToolDesc.Hidden gates both model visibility (runctx.Tools) and
/api/tools listing on runtime.fsTools, and registry.Execute re-checks
fail-closed. ToolExecCtx gains OwnerID/WorkspaceRoot/SettingBool.
go test ./... green (fs negative matrix, SSRF negatives, endpoint
fail-closed, cache).
Impact: link previews work end to end once the frontend wires the
fetcher; fs tools are inert until runtime.fsTools is enabled.
Refs: task feature/file-preview-caps; core 4d29301.
Why: agents rediscovered Gitea hosts and credential locations from
scratch each session, and the retired git-panel host kept failing with
TLS broken-pipe errors before the right host was recalled.
Evidence: curl git-home...:30319/api/v1/version with token from
~/.secrets/gitea -> {"version":"1.27.3"}; ssh -p 31411 git@git-ssh...
(BatchMode) -> "Hi there, weichen!" with key ~/.ssh/home_gitea (ED25519,
SHA256:GYXaOXIMeloQ1fKSfdJ/ndgPQzs/Ix7Hk62ekfkfQ20). Same section already
merged in Agent-System-Dev (PR #73).
Impact: AGENTS.md only; records hosts and credential locations, no
secret values (baseline hard boundary 2).
Refs: Agent-System-Dev PR #73
Why: generated deliverables (e.g. meeting-minutes docx) had no reliable
download path — the content endpoint set no Content-Disposition, so the
filename only survived via the front-end <a download> attribute, and
contentTypeOf returned text/plain for any artifact whose MIME column was
empty, mangling office/pdf uploads.
Evidence: artifactContent now always sets Content-Disposition (inline by
default, attachment with ?download=1) with an RFC 5987 filename* for
non-ASCII names plus an ASCII fallback; contentTypeOf consults an
explicit office-format extension table then mime.TypeByExtension before
the kind fallbacks. New tests cover disposition on both modes, the
encoded Chinese filename, and the extension/mime precedence chain;
go test ./... green.
Impact: front-ends can link direct download URLs and browsers receive
correct file types for docx/xlsx/pptx/pdf/csv/md regardless of the
stored MIME value.
Refs: task feature/chat-order-artifact-files.
Why: startup had only defaults and CLI flags; dev/test/prod operation needs
the standard baseline (explicit file, environment, explicit-only CLI) and a
read-only diagnostic.
Evidence: internal/config unit tests cover precedence, unset-CLI no-override,
presence-based file semantics, unknown-key/missing-file errors, env bool
parsing, explain source labelling and DSN redaction; full suite 12 packages
pass; live run of config explain with envs/prod.json shows correct layer
attribution and redaction.
Impact: serve gains --config and AGENT_SYSTEM_* env overrides (behavior
unchanged when neither is used); new config explain subcommand; envs/
examples and PROJECT.md configuration section.
Refs: documents docs/proposals/2026-09-30-configuration-surface.md; standards/configuration.md
Why: backend main pinned core 019617e (unfetchable post-migration) and the
store List* functions returned nil slices that encode as JSON null, which
crashed the frontend nav with .map on null (the 2026-09-30 white-screen).
Evidence: go test ./... across all packages; store regression
TestEmptyListsSerializeAsEmptyArrays asserts [] for empty project lists.
Impact: all 17 store list sites now initialize empty slices; build-blocker
core pin resolved; frontend no longer receives null for empty lists.
Refs: documents/records/2026-09-30-history-migration.md
Why: the composition adopted the delivery gate (Dev repo ef89fc8); resident
entry rules must carry the gate so agents load them without reading the
guidelines submodule first.
Evidence: text-only entry change; gate mechanics validated in the Dev repo
scratch-repo suite (19/19) and merged as PR weichen/Agent-System-Dev#1.
Impact: entry doc only; no code, contract, or build change in this repo.
Refs: Development-Guidelines 0.1.0-draft.7; Agent-System-Dev ef89fc8
Details (original message):
Add the vector RAG pipeline: paragraph chunker, OpenAI-compatible embeddings and rerank clients, dual-tier vector store with cosine search, and RRF hybrid retrieval
weichen