Why: depth 1 makes prev_sha and sha unrelated shallow roots, so update reports a fork instead of an upgrade.
Evidence: SELF the pin-fetch credential test; CI run 2829 showed the fork message.
Impact: test-pg fetches 200 commits from each pin, not the full history.
Refs: .task/feature-svc-d3-real-scenarios.md
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: test-pg only cloned the default branch, so prev_sha and legacy_sha were never built or attested.
Evidence: SELF python3 -m unittest tests.test_ci_credentials tests.test_lifecycle_pin (11 tests, exit 0).
Impact: the lifecycle suite prints the combination inventory and keeps login and rdm.ping as NOT_RUN.
Refs: .task/feature-svc-d3-real-scenarios.md
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: The live deploy journal never called fault injection, and test-lifecycle failed closed on login so CI could not require the real combination cases.
Evidence: SELF python3 -m unittest tests.test_svc_u2.U2ReleaseTests -q exit 0 (35 ok, 1 skipped) on 3602a87; receipt sha256 9b5785577751e8e9e7947cf9700fb8854b1c80142f34273ae25f3097ee4c9aa8.
Impact: test-pg builds the Dev binary and runs make test-lifecycle. Login and full-track deploy without SVC_RDM_ORIGIN stay NOT_RUN. No live data move and no account CLI.
Refs: recovery plan D2/D3; Dev #485
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: main gained the real-combination CI gate after the previous sync, and this branch still has to carry that gate with the isolated identity calls.
Evidence: SELF python3 -m unittest recovery/identity/u2/deploy/pin/isolation — 100 tests OK (2 skipped); bash -n scripts/svc.sh.
Impact: identity ensure/verify still pass DSN only through the environment; stop waits for the port and can keep the desired state. CI keeps the lifecycle pin step and the svc-realcombo job.
Refs: #485
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: an update that stops the old process and overwrites the binary cannot
get the previous version back, and a migration the old binary cannot read
must be refused before that stop. CI also had no fixed RDM SHA to build from.
Evidence: python3 -m unittest tests.test_svc_identity tests.test_svc_recover tests.test_svc_deploy_txn tests.test_lifecycle_pin passed. make test-lifecycle exits 2 when AGENT_PG_TEST_DSN is unset.
Impact: installs land in releases/<rid> and current switches with mv -T.
Deploy keeps the desired state and rolls back a non-terminal journal.
Schema compatFloor above the installed target is refused unless
--migrate-schema --confirm is paired with a backup. test-lifecycle fails
closed without PostgreSQL and an RDM checkout. Login stays NOT_RUN.
Refs: recovery plan r3.1 D2 and D3; Backend feature/d1-identity-open
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: on keyless runners the RDM cache clone and pins materialization
inside svc.sh fetched nested pins over raw SSH URLs (run 2608: host key
verification failed); the suite now injects the same URL-scoped token
header and insteadOf rewrite via environment when SVC_CI_GIT_* is set.
Evidence: local rerun stays green (env unset locally); CI run applies it.
Impact: tests + ci.yaml only; token travels via environment only.
Why: YAML env: values do not shell-expand, so $PWD-based binary paths
reached the suite as literal strings (run 2606: SVC_REAL_BIN not
executable).
Evidence: job log run 2606; paths now exported in the run script.
Impact: ci.yaml only.
Why: the realcombo steps unset the URL-scoped token env before
'submodule update', so nested pins fell back to raw SSH URLs and failed
host-key verification on the runner (first svc-realcombo run 2605).
Evidence: job log run 2605 shows ssh clone failures; reordered steps
keep insteadOf + token header active through the pins fetch.
Impact: ci.yaml ordering only.
Why: stub suites proved the script logic but not real artifact
behavior (the 2026-10-12 incidents passed all stub tests); the report's
U3 requires real-combination gates, and realcombo dry-runs exposed two
real gaps: fresh deploys closed identity governance one lifecycle late,
and pg-bound identity calls put the DSN on the command line.
Evidence: make test exit 0; local realcombo dry-run 16 OK (rows in the
task card); CI adds svc-realcombo (SQLite, both real binaries, full RDM
track) and a real-PG identity step on test-pg with env-delivered DSN.
Impact: tests/test_svc_realcombo.py (SVC_TEST_REAL gate, explicit
NOT_RUN otherwise); ci.yaml new job + test-pg step (RDM clone via
trusted-base URL-scoped token header, allowlist addition reviewed in
this PR); svc.sh delivers identity DSN via AGENT_SYSTEM_DSN and runs
manifest adoption in the same deploy that first registers svc.env;
modules/backend pinned to main (env fallback, backend #182).
Refs: .task/feature-svc-u3-realcombo.md; user goal directive 2026-10-12
(U3); backend #182; RDM #285.
Why: collaboration storage U7 (proposal sections 3/5/7) requires two-process
HTTP-level acceptance over shared PG, a reproducible load-test entry honoring
the load contract, and an auditable pre-release checklist; capacity windows and
dual-host runs need user-scheduled resources, so the deliverable here is the
verifiable entry, not capacity claims.
Evidence: make test green (composite); make test-multi-instance green locally
against an isolated PG (five scenarios incl. crash-takeover and restart
invariants); run_load.sh --profile smoke end-to-end green (SSE probe p95~18ms,
zero seq duplicates, stack auto-recycled); backend pinned to feature branch
with the three cross-instance fixes.
Impact: new tests/integration/multi-instance (CI test-pg job step), new
tests/load/collaboration (k6 + SSE probe + run_load.sh), release checklist at
docs/release-multi-instance.md; no runtime behavior change in Dev repo itself.
Refs: backend PR #164; task card feature/u7-load-fault-release.
Why: the job runs the full backend suite against real services; internal/runtime lint tests require python-docx (job 2290: ModuleNotFoundError 'docx') which the plain test job already installs.
Evidence: job 2290 log; all PG/Redis-dependent packages green in the same run (store 2.6s, pgbus, runmgr, ratelimit, server).
Impact: test-pg job setup only.
Refs: Dev PR #325.
Why: the runner-container guard had a doubled '[[' from a string replacement (exit '[: too many arguments'), and pgvector's extension is named 'vector' (control file vector.control) - CREATE EXTENSION pgvector failed with 'extension not available' (jobs 2285/2286 logs).
Evidence: job 2286 log 'ERROR: extension pgvector is not available ... pgvector.control: No such file'; package postgresql-16-pgvector 0.6.0 installed fine.
Impact: test-pg job only; repins backend to the #138 merge (30s SQLite busy timeout for serialized migrators).
Refs: backend PR #138; Dev PR #325.
Why: the runner container executes as root without sudo; the previous $SUDO -u postgres expanded to a bare '-u' and failed with exit 127 on the first test-pg run (job 2280 log).
Evidence: job 2280 log lines '/var/run/act/workflow/3: line 6: -u: command not found', exitcode 127; postgresql service itself started fine as root.
Impact: test-pg job setup steps only; also prints pinned pgvector/pg_trgm extension versions for acceptance records.
Refs: PR #325; docs/guides/team-storage.md section 5.
Why: the approved collaboration plan U2 requires CI to run the backend suite against real PostgreSQL+pgvector and Redis and fail when the services are missing, instead of silently skipping the tier.
Evidence: local gate green twice against PostgreSQL 18.6+pgvector and Redis 8.10.2 (isolated svc.sh instance); the target fails fast when AGENT_PG_TEST_DSN/AGENT_REDIS_TEST_ADDR are unset.
Impact: new make target test-backend-pg (env-guarded) and a required test-pg job (apt postgresql-16-pgvector + redis-server, ephemeral per-run services); existing test job unchanged.
Refs: proposals/2026-10-11-collaboration-storage-and-distribution.md U2; docs/guides/team-storage.md section 5.
Why: cleanup, component synchronization, receipt recording and credential handling could contradict the documented development boundaries.
Evidence: SELF 75 isolated governance tests and source checks are captured in .task/feature-governance-reliability.md; combination and CI verification follow.
Impact: hardens local task and CI tooling, aligns policy and component documentation, and adds continuous regression tests without changing product behavior.
Refs: .task/feature-governance-reliability.md; policy/governance.md; policy/platform-admission.md
Why: lint.py imports python-pptx and openpyxl at import time, so installing only python-docx still skipped the layout finding.
Evidence: actions run 680 failed TestLintErrorContinues with lint reason missing after python-docx was present.
Impact: CI installs python-docx, python-pptx, and openpyxl from the Tsinghua index.
Refs: feature/agent-stack-r0-r1
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: The runner image has Python but not python-docx, so TestLintErrorContinues failed after checkout succeeded.
Evidence: Agent-System-Dev actions job 1012 reported ModuleNotFoundError: No module named docx.
Impact: The package comes from the Tsinghua PyPI mirror. Library CI still stops until SUBMODULE_TOKEN is set.
Refs: feature/agent-stack-r0-r1
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: The runner cannot reach github.com, so actions/checkout and setup-node would time out before make test.
Evidence: Go 1.27.1 comes from mirrors.aliyun.com and Node 22.23.3 from npmmirror. Checkout uses the job token and GITHUB_SHA.
Impact: Submodule checkout still requires SUBMODULE_TOKEN. A missing secret fails the job with that message.
Refs: feature/agent-stack-r0-r1
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: make test downloads Go modules, and the runner cannot reach proxy.golang.org.
Evidence: The test job sets GOPROXY=https://goproxy.cn,direct. The step name lists boundaries, pins, libs, core, backend, frontend, ui, and agent-command.
Impact: npm still uses the default registry; add the npmmirror only if npm ci times out.
Refs: feature/agent-stack-r0-r1
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: the job token only authorizes the event repository; sibling private
submodules (Development-Guidelines et al.) returned 404 on the first run.
Evidence: run 194 log — clone of deps/development-guidelines failed with
repository not found despite the insteadOf rewrite.
Impact: ci.yaml reads secrets.SUBMODULE_TOKEN (stored via API, never in
tree or chat); no other change.
Refs: run 194; PR#11
Why: PR test evidence depended on the agent's word (single point); the
running binary could not answer what combination it was built from; and
irreversible operations lacked an explicit agent-stops-here boundary.
Evidence: Gitea runner online (home-runner, ubuntu-latest+docker); make
version prints describe+component SHAs; two-key line maps to baseline hard
boundary 1. CI first run observed on this PR before merge (see Actions tab).
Impact: .gitea/workflows/ci.yaml runs make test + make build on PR/push;
Makefile injects main.version (make version / agent-system version);
AGENTS.md and adoption register the two-key rule. No product code change.
Refs: policy/adoption.md; release-operations requirements
agent-exec-transport
weichenandCursor