Why: include the current integration baseline before verification.
Evidence: fetched origin/main=82c8989973112d0533ee7c70654c907269df7859.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
Why: include the current integration baseline before verification.
Evidence: fetched origin/main=84900d44d34d2d42428a5b5de1d92328d90a9950.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
Why: the S9 acceptance corrections left no repeatable harness for the
duty-flows browser verification, and the doc-level in-flight confirmation
path had never been driven in a real browser; the legacy e2e.sh seeding
predates the login-key credential contract.
Evidence: bash scripts/e2e-duties.sh end-to-end PASS — granted non-admin
preview 200, in-flight round + reviewer change surfaces the confirmation
checkbox and a confirmed save invalidates the doc submission, idempotent
re-save applied=0 unchanged=1, unauthorized member preview 400,
void-legacy then doc submit passes (evidence row in this card).
Impact: scripts-only; no product code or pins change.
Refs: S9 acceptance report P0-B
Why: Emergency Codex fallback must not run before Cursor, Antigravity, and PI have each been exhausted.
Evidence: Cursor transport 5e6c93f765c232783a7dccb05544e26c386fb031; 29 config tests, focused emergency controller test, and 449 governance tests passed with 19 skips.
Impact: Emergency-mode implementation, repair, and test roles now require exactly one enabled profile for each normal adapter while preserving configured order.
Refs: .task/feature-agent-cli-enforcement.md
Why: The approved contract requires exactly three safe attempts for Cursor, agy, and PI, then one Codex emergency attempt, never host implementation authority.
Evidence: Routing/config tests were RED for never, max attempts, forbidden adapters, and legacy host fallback; 133 related tests now pass with one opt-in probe skipped.
Impact: Normal emergency-mode roles reject all non-Cursor/agy/PI adapters, project fallback is never, and Codex remains available through emergency/review paths.
Refs: independent Codex review of feature/agent-cli-enforcement
Why: Success-shaped receipts and caller-writable history must not authorize apply/admit without controller-originated provenance.
Evidence: Fabricated, unsigned, tampered-history, and moved-ref tests were RED before HMAC binding; 52 controller/history/admission tests now pass.
Impact: Stores a local 0600 attestation key, signs successful history including transport commit SHA, and verifies it before apply/admit.
Refs: independent Codex review of feature/agent-cli-enforcement
Why: Safe routing must not retry unproven cleanup, ignore worker fail-stop terminals, or lose out-of-scope rename sources.
Evidence: New tests were RED for each Review finding; 44 controller and admission tests now pass with diff-check clean.
Impact: Tightens safety classification and path inventory only; routing order and provenance contract remain unchanged.
Refs: independent Codex review of feature/agent-cli-enforcement
Why: Agents must use configured CLIs for implementation and testing, while remaining free to inspect, plan, test, and iterate within bounded scope.
Evidence: Task-card contract tests and the full 430-test governance discovery pass with 19 existing skips; diff-check passes.
Impact: Defines three safe attempts per normal CLI, one emergency Codex attempt, fail-stop terminals, and mandatory apply/admit integration.
Refs: feature/agent-cli-enforcement
Why: Implementation changes must be mechanically tied to a successful CLI transport instead of relying on agents to follow prose.
Evidence: Admission tests went RED without apply/admit; 10 admission, 41 agent-exec, and 62 task tests now pass with exact staged-content matching.
Impact: Adds validated transport fast-forwarding and rejects unproven implementation commits while exempting task cards and approved plan/spec paths.
Refs: feature/agent-cli-enforcement
Why: Normal execution roles must exclude Codex while a single enabled Codex adapter remains available for emergency routing and review.
Evidence: Parser invalid-case tests were RED before validation; all 17 config tests and agent_exec validate now pass.
Impact: Enables three attempts per normal profile, stable Cursor routing, and a validated emergency Codex profile.
Refs: feature/agent-cli-enforcement
Why: Codex CLI is permitted only after every normal execution profile exhausts three safe attempts, with no host implementation fallback.
Evidence: The emergency RED test lacked the fourth Codex attempt before implementation; 32 controller/history executions now pass.
Impact: Adds one emergency attempt for implementation, repair, and test roles while preserving fail-stop statuses and review isolation.
Refs: feature/agent-cli-enforcement
Why: Each non-Codex execution profile must receive three safe attempts before routing advances.
Evidence: The focused RED test observed one retryme launch before implementation; the controller and history suites now pass 31 executions.
Impact: Retries only unavailable and failed-clean outcomes; existing fail-stop statuses remain immediate.
Refs: feature/agent-cli-enforcement
Why: Routing needs an explicit per-profile attempt limit and a named Codex emergency profile before controller enforcement can be implemented.
Evidence: The new parser test failed on unknown keys before implementation; all 15 config test executions now pass and git diff --check is clean.
Impact: Adds backward-compatible optional configuration fields only; routing behavior is unchanged in this commit.
Refs: feature/agent-cli-enforcement
Why: env ensure downloads an interpreter and wheels. The real-combination start test allows 120 seconds, and that download made rc1 start time out.
Evidence: CI job 5184 timed out in test_U3_real_agent_identity_lifecycle_sqlite on scripts/svc.sh rc1 start.
Impact: Start only mounts a base layer that already exists. The first build is an explicit env ensure, run outside start.
Refs: Agent-System-Dev pull 570
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: The dependency layer already existed in the backend, but service startup never passed it, so the sandbox could not import the packages the host already had.
Evidence: SELF go tests recorded on the backend commit. svc.sh only exports the env root after env ensure leaves a current base layer, and a failed download does not block startup.
Impact: Instances share runtime/envs. Existing exec= lines are left as stored. Proposal, backend, and frontend gitlinks move to this branch.
Refs: .task/feature-layered-python-env.md
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: include the current integration baseline before verification.
Evidence: fetched origin/main=dc86a023605d00fccca1b9679b4cdad96373bdee.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
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: PR #567 CI runs without a global Git user identity, causing every successful agent transport commit to fail with exit 128.
Evidence: The focused no-identity regression plus AgentExecControllerTests and AgentExecHistoryTests pass (30 executions total); git diff --check passes.
Impact: Supplies a deterministic non-secret identity only to git commit-tree; routing, receipts, and transport validation are unchanged.
Refs: #567
Why: include the current integration baseline before verification.
Evidence: fetched origin/main=919f75cf7b39ca357e212a08a0ba5ba3d9a79b51.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
Why: retain which CLI handled each task so routing can improve from real outcomes over time
Evidence: AgentExecHistoryTests and git diff --check
Impact: appends sanitized source-external JSONL records without changing run outcomes
Refs: feature/goal-first-delivery
Why: make the three approved CLIs the only active implementation and test routes
Evidence: adapter unit tests, configuration validation, diff check, and real marker probes for all three profiles
Impact: narrows project routing and fixes agy prompt binding for noninteractive execution
Refs: feature/goal-first-delivery
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: Hide provider-specific noninteractive, model, permission, workspace, prompt, and output differences behind one adapter seam.
Evidence: Cursor observed red then 22 green config/adapter tests; the host replays them and whitespace checks during integration.
Impact: Adds deterministic direct-argv adapters and probes without launching providers or enabling project routing.
Refs: docs/superpowers/plans/2026-10-15-project-agent-execution.md
Why: Establish one validated project-local routing document before any real Agent execution is enabled.
Evidence: Cursor observed red then 12 green configuration tests; the host replays those tests, the temporary validation target, and whitespace checks during integration.
Impact: Adds the U1 configuration contract and parser only; real routing, adapters, controller execution, and governance adoption remain disabled.
Refs: docs/superpowers/plans/2026-10-15-project-agent-execution.md
Why: Chat, lightweight, fallback and vision settings need a seed path, and the backend now sends current-turn images only to models declared capable of seeing them.
Evidence: The backend change is cfde38d. The seed test covers fallback and vision settings. A local image conversation on each builtin profile returned the pictured content from the expected model, and removing the chat model from the capable list switched that turn to the vision model.
Impact: New runs can use the four builtin profiles and image input. Secrets stay in the ignored env file. The live 8788 process is unchanged.
Refs: .task/feature-four-model-vision.md
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: Ensure implementation authorship is delegated per unit while the primary model remains responsible for planning and delivery management.
Evidence: The focused task-card generation regression went red then green, and git diff --check passed.
Impact: Every repository-writing implementation unit gets a fresh Cursor CLI run; unavailable or failed Cursor work is recorded as a blocker instead of being silently implemented by the primary model.
Refs: .task/feature-goal-first-delivery.md
Why: Keep authorized work focused on its stated outcome, prevent private scope expansion, and remove mechanical human stage gates.
Evidence: make test-governance passed 344 tests with 18 environment-dependent skips; receipt is recorded in the task card.
Impact: New tasks use goal-first execution while retaining real decision points, independent review, evidence, platform admission, and high-consequence dual keys.
Refs: .task/feature-goal-first-delivery.md
Why: deploy needs to choose registered plugins and ship them inside one agent-system, then refuse the release when they are not active.
Evidence: python3 -m unittest tests.test_svc_plugins tests.test_svc_user_help tests.test_svc tests.test_svc_u2 (SELF, 63 passed, 1 skipped). Real compose build and make test-lifecycle are NOT_RUN.
Impact: --with is mutually exclusive with --from rdm and --from release. Existing non-interactive deploys stay without plugins unless plugins= is already recorded. Changing the plugin id set on a live service requires --allow-source-switch.
Refs: .task/feature-svc-plugin-compose.md modules/documents/docs/proposals/2026-10-14-svc-plugin-compose-deploy.md
Co-authored-by: Cursor <cursoragent@cursor.com>
Why: card_merge only recognized "Merge ... from <branch> into main"
merge commits, so a PR landed as a squash (real case: #457) left its
task card in .task/ forever; task-close reported nothing to close and
precheck showed no debt. Closing required a manual backfill (PR #541).
Evidence: card_merge now returns sha|subject|pr and falls back to a
Gitea API lookup by head ref or label (deleted branches keep the
original name only in label), gated by merge_commit_sha ancestry in
origin/main; precheck [cards] reuses card_merge so both surfaces agree.
Live probe against production Gitea returns 226a4af15791|Reset
per-service runtime state in svc.sh env_config|457 for
feature/svc-pidcheck-leak; stub-server tests cover squash landing,
decoy skip and the not-on-main refusal; full governance suite follows
in-card evidence.
Impact: scripts/task.sh close and precheck detect squash-merged cards
when GITEA_TOKEN is set (merge-commit path unchanged without token);
tests/test_task.py adds two regressions; no protocol change for
existing merge-commit closures beyond the appended pr field.
Refs: card-merge-misses-squash-merges (memory); PR #541 manual closure.
Why: F2 token extraction predates the two-section (HOME/IMC) layout of
~/.secrets/gitea so it concatenated both tokens and always failed; F6
registered a grep BRE alternation that facts-verify unescapes into a
literal pipe, failing re-verification and misaligning precheck's naive
pipe split into a spurious F6 date-format expiry warning, while both
underlying facts still hold.
Evidence: make facts-verify now 6/6 PASS with last_verified refreshed
to 2026-10-14; new regression test for escaped pipes in facts rows;
full governance suite 342 tests OK (18 skipped); precheck [facts] line
reports all facts fresh on the fixed parser.
Impact: policy/facts.md F2 (HOME-section awk extraction) and F6
(negated pipe-free grep); scripts/task.sh precheck neutralizes escaped
pipes the same way facts-verify does; tests/test_task.py covers it.
Refs: precheck WARN 过期事实 F6(日期格式); AGENTS.md 证据与验证门禁.
Why: pipefail treated a recent merged-card match as a failed git log after grep exited early, hiding completed cards from precheck and task-close.
Evidence: A disposable long-history regression failed before the fix and passed after it; make test-governance passed 341 cases. Git history confirms PR #489/#488/#486 merge SHAs.
Impact: Detect merged cards reliably and archive only the three authorized trial cards; four unrelated eligible cards remain untouched.
Refs: Dev merge queue #518; Dev #489, #488, #486
Why: Finish Dev #488 with delivered Frontend and RDM objects, explicit audit grants in private browser fixtures, and owned process cleanup.
Evidence: Tool-imported SELF precheck, final sync, make test, and real-browser runs retain failures and passes; latest isolated runtime completed seven browser cases. Final exact-head GATE and independent parent Review follow.
Impact: Preserves other main components and existing CI jobs while the one final candidate is fixed for formal checks.
Refs: Dev PR #488; Frontend PR #73; RDM PR #301; queue Issue #518
Why: include the current integration baseline before verification.
Evidence: fetched origin/main=f00165e8d71e4dc8ba0ceb7dc50c1a8a3deb6d60.
Impact: integrate upstream changes without rewriting task history.
Refs: task-sync
Why: A LOCAL_READY receipt must permit normal target advancement without accepting a rewind or divergent target history after its observation.
Evidence: Two disposable real Git regressions reproduced erroneous successful recovery before the guard; all 18 task recovery tests pass afterward. Raw SELF receipts are in the task card.
Impact: Reject abnormal target ancestry before receipt import; preserve normal local recovery and strict delivery defaults. Correct stale implementation-stage documentation without activating the queue.
Refs: Documents implementation plan ba4044d U2; independent review R1 and R2; .task/feature-merge-queue-policy-draft.md
Why: agents need consistent entrypoints and recording defaults to avoid retrying synchronization and CI outside the final delivery window.
Evidence: SELF generated-task-card red-green regression and governance suite recorded through make evidence; pending shared record created as Issue #518.
Impact: rules, task template and capability matrix stay explicit about pending activation, default strict startup, manual coordination and unchanged CI triggers.
Refs: U4; policy/merge-queue.md; .task/feature-merge-queue-policy-draft.md
agent-exec-transport
weichen