chore(release): v3.11.5#555
Merged
Merged
Conversation
…ness chain, mincut, GOAP executor, and SONA Full remediation of docs/analysis/SYSTEM-INTEGRITY-AUDIT-2026-07-04.md's P0-P2 findings, tracked in docs/plans/SYSTEM-INTEGRITY-REMEDIATION-GOAP-PLAN-2026-07-04.md. Highlights: - Dream engine: real pattern creation on insight apply, error/outcome concept nodes from real ADR-110 data, and a follow-up fix ensuring the concept refresh actually runs on the real automatic path (not just via test wrappers); persistTaskOutcome now genuinely promotes pending insights instead of blindly incrementing an applied counter - Witness chain: Ed25519 signing wired via persistent key directory, archival/append-after-archival bug fixed, CI verification gate added - Mincut: real event-bus wiring (was a no-op stub) and scaffold-vertex detection fixes - GOAP: real domain-API dispatch executor replacing the always-mock one - SONA: EWC++ cold-start deadlock fixed via independent request-count persistence - ADR-110 pattern-nulls wired into the real production capture path - Daemon/snapshot visibility, TTL unit bugs, and 8 ADRs re-statused with additive verification notes Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
… vector count (#553) qe_patterns is the source of truth for pattern count and domain/tier breakdown; the RVF/HNSW vector index only covers patterns that have been embedded, so getStats() silently hid any pattern stored without a matching embedding (aqe learning stats reported the vector count as Total, and "By Domain" was hardcoded empty). Also fixes the rvf-pattern-store.test.ts Stats test to attach an isolated non-unified SQLite store per test — needed once getStats() started reading real SQLite data, since the auto-attached ADR-046 unified store is a process-wide singleton shared across test files. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…otonic stats Three compounding bugs in the recordOutcome()/storePattern() path, found while chasing a flaky qe_pattern_usage-count test: - recordOutcome() wrote pattern usage twice per outcome: once via patternStore.recordUsage() (which already persists to the wired SQLitePatternStore) and again via an explicit direct SQLite call added only to carry metrics/feedback (which the interface didn't accept). Threaded metrics/feedback through IPatternStore.recordUsage() instead and removed the redundant write — every outcome was inflating qe_pattern_usage row counts and qe_patterns.usage_count by 2x. - PatternStore.store() (in-memory backend) never applied the #447 fix that RvfPatternStore already had: when ON CONFLICT(name, qe_domain, pattern_type) preserves an existing row's id, the in-memory indices kept using the discarded id, so every later recordUsage()/promote()/ delete() call silently failed against SQLite with "Pattern not found". - SQLitePatternStore.storePattern() resolved the ON CONFLICT-preserved id AFTER the insert transaction, but inserted the embedding using the discarded id INSIDE the transaction — violating the qe_pattern_embeddings -> qe_patterns(id) foreign key whenever a pattern name collided with an existing row. Now resolves the actual id before the embedding insert, in the same transaction. - QEReasoningBank.getStats() only consulted the DB aggregate when the in-memory counter was exactly zero, so reported totals regressed the moment in-memory ticked past zero mid-session (a cold-start call correctly showed the historical DB total; the next call after one recordOutcome() dropped to reporting "1"). Since qe_pattern_usage is written synchronously, the DB aggregate is always >= the in-memory counter — now takes the max of the two, keeping the number monotonic. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…k remaining work - Backup-verified cleanup of .agentic-qe/memory.db: 200/200 degenerate mincut_snapshots rows, 501/501 degenerate mincut_history rows, 530/530 test-fixture execution_results rows, 705/705 test-fixture executed_steps rows (all confirmed 100% junk before deleting); corrected 72 dream_insights.applied rows corrupted by the now-fixed counter bug back to the correct boolean value. Full backup + before/after row counts + integrity check per data-protection policy. - Fixed 4 unused-vars lint errors in goap-planner.ts/goap-execute.ts (2 pre-existing + 1 that surfaced after removing the first + 1 in goap-execute.ts). Left two genuinely unimplemented stub params (updateActionStats duration tracking, findSimilarPlan fuzzy matching) honestly marked rather than scope-expanded. - Filed #554 to track everything from the system-integrity remediation that's deliberately deferred, so it isn't silently dropped. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Bumps package.json + package-lock + skills-manifest fleetVersion to 3.11.5. Also fixes .claude-plugin/plugin.json and assets/skills/skills-manifest.json fleetVersion, which had drifted stale at 3.11.0 across the last 4 releases (found via full-repo version-string audit).
Skill Validation Report
Only Tier 3 failures block merge |
Test Suite MetricsCI Test MetricsDate: 2026-07-07 07:52:39 UTC Current State
Progress from Baseline
Generated by Optimized CI |
MCP Tools Test SummaryValidation Results
Unit Test Results
Other Results
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
System-integrity sweep of the self-learning loop plus pattern-usage stat fixes: dream-cycle insights now genuinely become reusable patterns (a gate bug meant this only ever ran in manual tests, not real scheduled cycles), pattern usage/stats bugs are fixed (double-write, id-desync lookup failures, mid-session regression), and the witness-chain, agent-topology (mincut), GOAP execution, and SONA continual-learning subsystems each had real dormant-code/data-integrity bugs fixed. Full details: CHANGELOG, audit, remediation plan.
Verification
npm run build— clean, all 3 bundles (CLI/index/MCP) produced.npm run typecheck— 0 errors.npx vitest run tests/unit/— 17388 passed, 13 skipped, 0 failed (628/629 files, 1 pre-existing skip).npm run performance:gate— 15/15 gates passed.npm run test:regression(test:unit:fast && test:unit:heavy && test:unit:mcp) — passed clean on a standalone re-run (7026 passed | 2 skipped, 240/240 files). One transient failure appeared only when this suite ran immediately after the fulltests/unit/sweep above, sharing a hardcoded (non-PID-unique)/tmppath insona-persistence.test.ts; re-running standalone reproduced 0 failures, confirming it was a same-path collision between two back-to-back local runs, not a real regression.npm run test:ci— 8572 passed | 10 skipped, 2 pre-existing failures (see below), neither touched by this branch's commits.aqe init --autoin a fresh temp project — completed cleanly (85 skills, 60 agents, hooks/MCP/workers configured).aqe --version→3.11.5;aqe status→ healthy;aqe learning stats→ real per-domain breakdown (confirms the pattern-store stats fix);aqe agent list/aqe health→ both respond cleanly (health correctly reports the background daemon as "not configured" rather than falsely healthy).npm pack→ freshnpm installin an empty dir →node node_modules/.bin/aqe --version→3.11.5, exit 0, 0 vulnerabilities..claude-plugin/plugin.jsonandassets/skills/skills-manifest.jsonhad drifted stale at3.11.0across the last 4 releases; fixed alongside this bump.PRAGMA integrity_check= ok): removed 200/501 degenerate mincut placeholder rows, 530/705 dead GOAP test-fixture rows, and corrected 72 corrupteddream_insights.appliedcounter values back to boolean1.Failure modes
test:cifailures are pre-existing, not introduced by this branch — verified by checking out unmodifiedorigin/mainin an isolatedgit worktreeand reproducing both identically:tests/integration/llm/multi-provider.test.ts— 2 assertions expect stale hardcoded Claude model-ID/display-name mappings (claude-opus-4-7/Claude Sonnet 4) that don't match this environment's actual model registry; fails identically onorigin/main.tests/integration/ruvector/native-hnsw-real-fixture.test.ts— expects a hardcoded recall count (1600) against this project's real, organically-growing.agentic-qe/memory.db; reproduced the exact same failure (1397) on unmodifiedorigin/mainwhen pointed at the same real DB (it only skips on main with a fresh/empty DB). Test design depends on live DB content rather than a frozen fixture.sona-persistence.test.tstransient failure described above only reproduces when two full local test suites are run back-to-back against the same hardcoded/tmp/agentic-qe-test-sona-persistencepath (not a realistic CI scenario, which runs the suite once); the underlying A10 fix passes 100% reliably in isolation and in its normal composite run.dream_insights.appliedrows' history before the fix, nor several other deliberately-deferred items (hook-trigger breadth, 13/40 stub GOAP actions, dormant mincut event handlers, unsigned historical witness-chain rows) — all tracked in #554.Required check (issue #401)
Optional context
npm-publish.yml/tests/fixtures/init-corpus/: no./tests/fixtures/init-corpus/run-gate.shlocally? not applicable