test(checkers): deepen observable behavior coverage

.agent-loop/initiatives/WS-QUAL-001-backend-coverage-floor/CHUNK_MAP.md

@@ -18,9 +18,9 @@ ladder are superseded before implementation.
 
 | Chunk | Purpose | Risk | State |
 |---|---|---:|---|
-| `WS-QUAL-001-PLAN2` | Reconcile current hosted baseline, retire obsolete machinery, and define the small closure sequence | L1 | Planning in progress |
-| `WS-QUAL-001-02R` | Project/setup observable behavior coverage | L2 | Proposed after PLAN2 |
-| `WS-QUAL-001-03R` | Checker observable behavior coverage | L2 | Proposed after 02R only if more headroom is needed |
+| `WS-QUAL-001-PLAN2` | Reconcile current hosted baseline, retire obsolete machinery, and define the small closure sequence | L1 | Merged PR #260 |
+| `WS-QUAL-001-02R` | Project/setup observable behavior coverage | L2 | Merged PR #265 |
+| `WS-QUAL-001-03R` | Checker observable behavior coverage | L2 | Implementation in progress |
 | `WS-QUAL-001-04R` | Change the exact global hosted CI floor from 78 to 90 after current-main proof | L1 | Proposed after measured >=90.25% proof |
 
 One chunk maps to one PR. A test chunk may close early when its behavioral scope

.agent-loop/initiatives/WS-QUAL-001-backend-coverage-floor/DISCOVERY.md

@@ -2,6 +2,27 @@
 
 ## Audited baseline
 
+### Current-main refresh for 03R
+
+Backend run `30921410531` on current-main merge commit `5b853d50` completed
+3,068 tests, covered 21,453 of 23,938 statements (89.619016 percent), recorded
+727.166 seconds total hosted wall time, and a 567.994-second slowest lane.
+Reaching 90.25 percent on this denominator requires 21,605 covered statements,
+a net gain of 152. The focused 03R test union covers 168 statements missing
+from this hosted report: checker service 107, runner 45, and compiler 16. That
+projects 21,621 / 23,938, or 90.320829 percent; hosted exact-head fan-in remains
+authoritative.
+
+Checker-owned gaps are sufficient and remain unchanged by ART: service 169,
+runner 45, compiler 26, router 12, repository 11, gate queue 2, and pre-review
+gate 1. Existing checker tests are integration-heavy. Direct fast coverage is
+still missing for observable policy-shape rejection, registry ordering and
+conflicts, routing priority, blocking-policy escalation, role-sensitive result
+redaction, and bounded gate recovery outcomes. These are the preferred 03R
+test seams; unrelated TASK or ART lifecycle paths remain out of scope.
+
+### PLAN2 historical baseline
+
 Hosted Backend run `30854931616` on the final PR #249 tested tree
 `19d48f7ea4bf20cb29f03cbba54f98683ce52661` produced:
 

.agent-loop/initiatives/WS-QUAL-001-backend-coverage-floor/STATUS.md

@@ -2,29 +2,37 @@
 
 ## Current state
 
-`WS-QUAL-001-PLAN2` merged through PR #260. `WS-QUAL-001-02R` is now at its
-external-review gate with a test-only implementation and passing required
-internal reviews.
-
-The latest complete current-main hosted baseline is Backend run `30891776021`
-on `b47a7e64`: 88.603709 percent across 23,455 statements and 2,936 tests.
-02R's first exact-head hosted run passed but measured 21,003 covered
-statements, or 89.545939 percent, one statement below the chunk's 89.55-percent
-target. A meaningful project-agent composition test now covers that remaining
-factory statement and awaits exact-head hosted verification. The global CI
-floor remains 78 percent; named protected subsystem checks remain blocking at
-90 percent.
+`WS-QUAL-001-PLAN2` merged through PR #260. `WS-QUAL-001-02R` merged through
+PR #265 with all 2,996 exact-head tests passing and 21,004 / 23,455 statements
+covered (89.550203 percent).
+
+The current-main baseline after ART PR #268 is Backend run `30921410531` on
+`5b853d50`: 3,068 tests completed, 21,453 / 23,938 statements covered
+(89.619016 percent), 727.166 seconds hosted wall time, and a 567.994-second
+slowest lane. The global CI floor remains 78 percent; named protected subsystem
+checks remain blocking at 90 percent.
 
 Historical QUAL work delivered the isolated database runner and test-integrity
 guards through PRs #103, #105, and #108. The many stopped semantic-analysis
 replacements remain historical evidence, not work to resume.
 
 ## Current gate
 
-Require Agent Gates, CodeRabbit, all Backend semantic lanes, final coverage
-fan-in, and human review for `WS-QUAL-001-02R`. Hosted coverage must be at least
-89.55 percent, and an unexplained hosted runtime increase above 10 percent
-stops merge readiness.
+`WS-QUAL-001-03R` is the current implementation chunk. It must add meaningful
+checker-owned behavior tests and gain at least 152 covered statements on the
+current-main denominator to reach the 90.25-percent headroom target. The
+focused test union measures
+168 unique previously missing checker statements and projects 21,621 / 23,938,
+or 90.320829 percent. Require Agent
+Gates, CodeRabbit, all Backend semantic lanes, final coverage fan-in, six
+internal reviewer tracks, and human review. An unexplained focused or hosted
+runtime increase above 10 percent stops merge readiness.
+
+The 94-case focused implementation selection passes in 37.76 seconds with
+narrow checker-module coverage. The isolated full `test_checkers.py` run
+reached the 1,200-second local ceiling after approximately 95 percent completion
+with 163 passing tests and no failure output. It is not a complete pass; hosted
+Backend remains mandatory.
 
 ## Stop condition
 

.agent-loop/initiatives/WS-QUAL-001-backend-coverage-floor/chunks/WS-QUAL-001-03R-checker-behavior-coverage.md

@@ -5,6 +5,12 @@ Parent initiative: `WS-QUAL-001`
 Goal: if still necessary after 02R, add meaningful fast tests for current
 checker service, repository, runner, compiler, and routing gaps.
 
+Current-main basis: Backend run `30921410531` on merge commit `5b853d50`
+completed 3,068 tests with 21,453 / 23,938 covered statements (89.619016
+percent), 727.166 seconds hosted wall time, and a 567.994-second slowest lane.
+The 90.25-percent target requires 21,605 covered statements, a net gain of 152
+on this exact denominator.
+
 Risk: L2 test-only; P2.
 
 Allowed files:
@@ -16,17 +22,23 @@ Not allowed: production code, migrations, workflow/threshold changes, skips,
 xfails, coverage exclusions, assertion deletion, or duplicate system flows.
 
 Acceptance: observable outcomes, focused proof, recorded runtime impact,
-complete hosted Backend, and exact global coverage of at least 90.25 percent.
+complete hosted Backend, and exact global coverage of at least 90.25 percent
+(21,605 / 23,938 on the starting denominator).
 If meaningful checker gaps are exhausted first, stop and plan one explicit
 owner-specific successor rather than add artificial tests.
 
 Verification commands:
 
 - `cd backend && ruff check tests/test_checkers.py`
-- `cd backend && WORKSTREAM_TEST_ADMIN_DATABASE_URL=postgresql+asyncpg://workstream:workstream@localhost:5433/postgres python scripts/run_isolated_tests.py --metadata-json .ci/qual-03r-database.json --timeout-seconds 1200 -- pytest -q tests/test_checkers.py`
+- `cd backend && qual_tmp=$(mktemp -d /tmp/ws-qual-03r-db.XXXXXX) && WORKSTREAM_TEST_ADMIN_DATABASE_URL=postgresql+asyncpg://workstream:workstream@localhost:5433/postgres .venv/bin/python scripts/run_isolated_tests.py --metadata-json "$qual_tmp/metadata.json" --timeout-seconds 1200 -- .venv/bin/pytest -q tests/test_checkers.py`
 - Hosted Backend semantic lanes and final coverage fan-in.
 - Test-delta scan for skips, xfails, deleted assertions, and coverage narrowing.
 
+The 1,200-second local full-file timeout is a bounded diagnostic limit on the
+constrained development machine and may be reached before collection completes.
+Only the hosted Backend exact-head run is authoritative proof of the complete
+suite pass.
+
 Required reviewers: senior, QA, test-delta, CI integrity, product/ops, reuse.
 Human focus: checker invariant ownership and fast-layer placement. This chunk
 does not change or test TASK-owned lifecycle behavior merely for percentage.

.agent-loop/initiatives/WS-QUAL-001-backend-coverage-floor/reviews/WS-QUAL-001-03R-external-review-response.md

@@ -0,0 +1,30 @@
+# WS-QUAL-001-03R External Review Response
+
+## Comments addressed
+
+- CodeRabbit reported one trivial documentation ambiguity: the contract listed
+  the 1,200-second local isolated-run bound without explaining that this
+  constrained machine may reach the bound before the full file completes.
+  The contract now keeps the same command and explicitly names hosted Backend
+  on the exact PR head as the authoritative complete-pass proof.
+
+## Comments deferred
+
+None.
+
+## Human decisions needed
+
+None for the external finding. Human merge approval remains required after all
+exact-head checks pass.
+
+## Commands rerun
+
+- `git diff --check`
+- `python3 scripts/check_markdown_links.py`
+- `PYTEST_DISABLE_PLUGIN_AUTOLOAD=1 python3 -m pytest -q scripts/test_lightweight_agent_gates.py`
+
+## Remaining risks
+
+- Backend semantic lanes and final fan-in must pass on the repaired exact head.
+- Hosted coverage must reach at least 90.25 percent and runtime must be compared
+  with current-main run `30921410531`.

.agent-loop/initiatives/WS-QUAL-001-backend-coverage-floor/reviews/WS-QUAL-001-03R-internal-review-evidence.md

@@ -0,0 +1,76 @@
+# WS-QUAL-001-03R Internal Review Evidence
+
+## Reviewed revision
+
+- Code SHA: `23235d506d9f16cb000ba9e9219e4f3e28ddbb79`
+- Reviewed base SHA: `cda59fc32e3a60e1e2c337cae7ebeaa94b95e12b`
+- Reconciled current-main SHA: `5b853d505b55b6465a472d25aa33caec4be18073`
+- Reviewed at: `2026-08-04`
+- Open reviewer sessions: none
+- Valid findings addressed: yes
+
+## Circuit-breaker decision
+
+PASS with a documented size exception. The diff is large because 94 focused
+cases are added to one checker-owned test module, but it crosses one production
+boundary, changes no production code, and adds no workflow, migration, public
+contract, dependency, skip, xfail, or coverage exclusion. Splitting the same
+checker behavior across dependent PRs would add coordination cost without
+making the underlying boundary smaller.
+
+## Reviewer results
+
+| Reviewer | Result | Blocking findings | Notes |
+|---|---:|---|---|
+| senior engineering | PASS | None | Helpers and fakes remain reviewable and assert concrete service outcomes. |
+| QA/test | PASS WITH LOW RISKS | None | Hosted exact-head Backend remains the authoritative acceptance proof. |
+| test delta | PASS | None | Additive behavior tests; no weakening, skips, xfails, or exclusions. |
+| CI integrity | PASS WITH LOW RISKS | None | No CI changes; coverage arithmetic is coherent and hosted proof remains required. |
+| product/ops | PASS | None | Checker routing remains distinct from human product review decisions. |
+| reuse/dedup | PASS | None | Added tests are fast boundary tests and do not duplicate the existing HTTP/PostgreSQL suffix. |
+
+Security/auth, architecture, and docs review are not required for this
+test-only chunk: it changes no production boundary, authorization behavior,
+architecture, public contract, or contributor documentation.
+
+## Findings reconciled
+
+The first reuse pass attributed pre-existing broad API tests to this chunk by
+using their post-insertion line numbers. Git history showed that
+`test_chunk10_checker_trial_runs_sample_submissions_through_real_api` and the
+other cited system tests already existed on `origin/main`; the 03R insertion
+only shifted them downward. Re-review of the actual insertion hunk found no
+duplicate system flow and withdrew the finding.
+
+During review, ART PR #268 advanced `main`. It touched ART production, tests,
+lane routing, and ART documentation but did not overlap the checker test or
+QUAL initiative delta. The branch merged exact current main without conflict;
+`git diff origin/main...HEAD` retains the reviewed 03R implementation. Main's
+new exact hosted baseline is recorded below and the final reconciled head
+receives focused reviewer confirmation before publication.
+
+## Commands and evidence
+
+- `cd backend && .venv/bin/ruff check tests/test_checkers.py` — pass.
+- Focused new-test selection — 94 passed, 76 deselected in 37.76 seconds.
+- `pytest --collect-only -q tests/test_checkers.py` — 170 tests collected.
+- Narrow coverage comparison against current main — 168 uniquely covered
+  statements: checker service 107, runner 45, compiler 16.
+- `git diff --check origin/main...HEAD` — pass.
+- Lightweight Agent Gates — 10 passed.
+- Allowed-path review — checker tests and the QUAL initiative tree only.
+- Test-delta scan — no skip, xfail, deleted assertion, coverage exclusion, or
+  CI bypass.
+- Full isolated `test_checkers.py` — the constrained local runner reached
+  approximately 95 percent (about 163 tests) without a failure before its
+  1,200-second bound. This is not recorded as a complete pass.
+
+## Remaining risks
+
+- Hosted Backend must pass every semantic lane and final fan-in on the exact PR
+  head.
+- Hosted fan-in must prove at least 21,605 / 23,938 covered statements (90.25
+  percent). The local projection is 21,621 / 23,938 (90.320829 percent).
+- Compare hosted wall and slowest-lane time with main run `30921410531`
+  (727.166 seconds wall; 567.994 seconds slowest lane). An unexplained increase
+  above 10 percent stops merge readiness.

.agent-loop/initiatives/WS-QUAL-001-backend-coverage-floor/reviews/WS-QUAL-001-03R-pr-trust-bundle.md

@@ -0,0 +1,115 @@
+# WS-QUAL-001-03R PR Trust Bundle
+
+## Chunk
+
+`WS-QUAL-001-03R` — Checker Behavior Coverage.
+
+## Goal
+
+Raise global backend coverage from the exact current-main 89.619016-percent
+baseline to at least 90.25 percent through meaningful checker-owned behavior
+tests, without changing production behavior or CI policy.
+
+## Human-approved intent
+
+Continue the approved QUAL initiative one bounded chunk at a time, preserve
+real PostgreSQL/system proof, and move fast business-rule coverage into the
+smallest owning layer rather than adding arbitrary infrastructure or shards.
+
+## What changed and why
+
+- Added 94 focused cases in `backend/tests/test_checkers.py` for runner,
+  compiler, service, authorization, routing, redaction, policy-lock, automatic
+  gate, recovery, provenance, and fail-closed behavior.
+- Updated only the QUAL initiative's discovery, contract, map, status, and
+  review evidence so durable memory matches the implementation.
+- The tests cover 168 statements that current main does not cover: checker
+  service 107, runner 45, and compiler 16.
+
+## Design chosen
+
+Use direct runner/compiler/service tests and bounded repository/session fakes
+for checker-owned decisions. Keep existing broad HTTP/PostgreSQL flows as the
+small contract suite and leave production composition unchanged.
+
+## Alternatives rejected
+
+- More arbitrary CI shards: they move runtime rather than establish ownership.
+- Duplicate end-to-end flows: slower and disallowed by the contract.
+- Artificial execution or coverage exclusions: they would weaken trust.
+- Production refactoring in this chunk: outside the reviewed test-only scope.
+
+## Scope control
+
+The implementation is limited to `backend/tests/test_checkers.py` and the
+`WS-QUAL-001` initiative tree. It changes no production code, migration,
+workflow, threshold, dependency, public API, skip, xfail, or coverage omit
+rule. The circuit breaker passed with a size exception because this remains one
+test file and one production boundary with no production behavior change.
+
+## Product behavior
+
+No product behavior changes. The tests preserve the distinction between
+checker routing recommendations and the only stored human review decisions:
+`accept`, `needs_revision`, and `reject`.
+
+## Acceptance criteria proof
+
+Current-main Backend run `30921410531` on `5b853d50` recorded 3,068 completed
+tests, 21,453 / 23,938 covered statements (89.619016 percent), 727.166 seconds
+hosted wall time, and a 567.994-second slowest lane. The target requires 21,605
+covered statements. Local differential coverage projects 21,621 / 23,938
+(90.320829 percent); hosted exact-head fan-in is authoritative.
+
+## Tests and checks run
+
+- Ruff: pass.
+- Focused selection: 94 passed, 76 deselected in 37.76 seconds.
+- Checker collection: 170 tests collected.
+- Git whitespace and allowed-path checks: pass.
+- Lightweight Agent Gates: 10 passed.
+- Local full checker file: approximately 95 percent completed without failure
+  before the 1,200-second runner bound; not claimed as a complete pass.
+
+## Test delta and CI integrity
+
+Tests are additive. No assertion was deleted or weakened, and no test was
+skipped, xfailed, deselected by CI, or excluded from coverage. No workflow,
+runner, lane, service, coverage command, or threshold changed.
+
+## Reviewer results
+
+Senior engineering, QA/test, test delta, CI integrity, product/ops, and
+reuse/dedup reviewed exact code SHA `23235d506d9f16cb000ba9e9219e4f3e28ddbb79`.
+All pass. The first reuse finding was withdrawn after git history proved the
+cited broad system tests were pre-existing main content shifted by the new
+insertion.
+
+## External review
+
+Agent Gates passed on the published head. CodeRabbit's one trivial finding was
+valid: the contract needed to distinguish the bounded local 1,200-second
+diagnostic timeout from hosted complete-pass authority. The contract now says
+that explicitly. Backend exact-head semantic lanes and final fan-in are still
+running; human review remains pending.
+
+## Remaining risks and follow-up work
+
+- Hosted coverage may differ from the local differential projection.
+- Hosted runtime must remain within 10 percent of the recorded baseline unless
+  the change is explained before merge.
+- Raising the blocking global floor is a separate successor decision after
+  this exact-head coverage is proven; it is not part of 03R.
+
+## Human review focus
+
+- Confirm tests exercise observable checker invariants rather than percentage
+  padding.
+- Inspect fake fidelity around policy provenance and gate state transitions.
+- Confirm hosted coverage is at least 90.25 percent and runtime remains within
+  the contract bound.
+
+## Human merge ownership
+
+This chunk stops after exact-head external checks and human review. It will not
+be merged without the user's explicit approval for this PR.