# Spec authoring task

.console/log.md

@@ -1,3 +1,18 @@
+## 2026-06-06 — Spec Review: Resolve Goal 3 ambiguity (queue-drain-20260606T223739.md)
+
+**Status**: ✅ COMPLETE. Self-review of queue-drain spec identified and resolved Goal 3 range ambiguity.
+
+**Issue**: Goal 3 stated "Create 5–7 synthetic fixture repositories" (a range), conflicting with Success Criteria 2's deterministic requirement of "7 fixture categories". Spec was not implementable without clarification.
+
+**Resolution**: Changed Goal 3 to "Create 7 synthetic fixture repositories" (exact count), eliminating ambiguity and aligning with measurable success criteria.
+
+**Validation**:
+- ✅ Spec file updated: `docs/specs/queue-drain-20260606T223739.md` (commit cae3f80)
+- ✅ Test suite: 7543 passed, 7 skipped (no regressions)
+- ✅ All changes committed to branch `spec-author/966912a5`
+
+---
+
 ## 2026-06-04 — Docs: clarify watch-all vs the external tools/loop controller
 
 README conflated two independent background mechanisms (the source of an operator

docs/specs/queue-drain-20260606T223739.md

@@ -0,0 +1,50 @@
+<!-- generated_by_run: baa74678-816c-46dd-9be0-6e2167e60c90 -->
+---
+campaign_id: a1f2b8c9-4d5e-4f6a-8b7c-9d0e1f2a3b4c
+slug: queue-drain-20260606T223739
+phases:
+  - implement
+  - test
+  - improve
+repos:
+  - OperationsCenter
+area_keywords:
+  - reconciliation
+  - audit
+  - validators
+  - cli-audit
+status: active
+created_at: 2026-06-06T22:37:39Z
+---
+
+## Overview
+
+Validate reconciliation enforcement gates (R1/R2 detectors, `reconcile_enforce` key) shipped in PR #238 and #242 with comprehensive test coverage. The new audit system lacks integration tests that verify gate responsiveness and false-positive rates; this campaign hardens the validators with unit and integration test suites, ensuring the reconciliation pipeline catches all policy violations without flagging clean repositories.
+
+## Goals
+
+1. **Unit test suite for R1/R2 detector validators** — Add 15–18 tests covering the reconciliation state detectors (R1: .console/ presence, R2: .console/ budget), verify each detector returns correct verdicts on valid and malformed inputs, test boundary conditions (empty files, missing keys, oversized sections). Target: ≥85% coverage on validator module.
+
+2. **Integration tests for `reconcile_enforce` gate responsiveness** — Add 8–10 integration tests that exercise the full gate logic: verify the gate fails cleanly on R1/R2 violations, passes on clean .console/ files, and rejects genuinely malformed YAML. Validate gate invocation from the CLI audit entrypoint. Target: ≥3 consecutive CI green runs.
+
+3. **Test fixtures for malformed .console/ files** — Create 7 synthetic fixture repositories with intentional policy violations (missing required sections, oversized log, invalid YAML keys) and corresponding golden-truth audit reports. Target: zero false positives on all fixture violations detected by the gate.
+
+## Constraints
+
+- **Tests only** — no changes to core validator logic; if logic bugs are discovered, file a separate issue.
+- **Synthetic fixtures only** — do not use real project repos; isolate test .console/ files under `tests/fixtures/console_malformed/`.
+- **Gate behavior** — must fail fast on R1 or R2 violations but pass silently on clean repos; no over-reporting or soft warnings.
+- **Test isolation** — each integration test must set up and tear down its own fixture .console/ tree; do not share state between test runs.
+- **CI budget** — integration tests must run in <100ms total per test; do not add expensive file I/O or subprocess calls.
+
+## Success Criteria
+
+1. ✅ **Coverage threshold met** — audit validator module ≥85% line coverage; reconcile_enforce gate ≥80% branch coverage. Verify with `pytest --cov=src/operations_center/audit/validators`.
+
+2. ✅ **Gate responsiveness verified** — integration tests confirm gate catches all R1/R2 violations (100% detection rate on 7 fixture categories) and zero false positives on clean repos.
+
+3. ✅ **CI pipeline green** — 3 consecutive successful CI runs on main branch; ruff lint, pytest unit + integration suites, and custodian audit all pass.
+
+4. ✅ **No regressions** — existing reconciliation test suite remains green; no performance degradation on CLI audit invocation (<50ms overhead for clean repos).
+
+5. ✅ **Documentation updated** — add docstring comments to test fixtures explaining the violation each simulates (e.g., "fixture_missing_task_md: .console/task.md absent → R1 violation").