agentic-qe 3.2.3 → 3.3.1

package/.claude/agents/v3/qe-accessibility-auditor.md

@@ -30,6 +30,11 @@ Working:
 - Extended aria-describedby descriptions for screen readers
 - Copy-paste ready code fixes for all violations
 - Comprehensive HTML/Markdown report generation
+- **EN 301 549 V3.2.1 EU compliance mapping** (harmonized European standard)
+- **EU Accessibility Act (Directive 2019/882) validation**
+- **WCAG-to-EN 301 549 clause mapping** (all 50+ web clauses)
+- **EAA product category validation** (e-commerce, banking, transport, etc.)
+- **Certification-ready EU compliance reports**
 
 Partial:
 - Screen reader testing (NVDA, VoiceOver, JAWS)
@@ -92,6 +97,21 @@ Use up to 6 concurrent auditors for large sites.
   - Spatial relationships and all visible text
 - **Context-Aware ARIA**: Intelligent label generation based on element semantics and user flow
 - **Developer-Ready Output**: Copy-paste code snippets for every violation found
+- **EN 301 549 Compliance**: Full mapping to European ICT accessibility standard V3.2.1:
+  - Chapter 9 Web content (50+ clauses mapped to WCAG)
+  - Automated, manual, and hybrid test method classification
+  - Clause-by-clause compliance reporting
+  - Remediation prioritization by test method
+- **EU Accessibility Act Validation**: Directive 2019/882 compliance checking:
+  - Product category validation (e-commerce, banking, transport, e-books, etc.)
+  - Requirements mapping to EN 301 549 clauses
+  - Exemption tracking (micro-enterprise, disproportionate burden)
+  - Annex I functional requirements validation
+- **EU Certification Reports**: Generate certification-ready compliance documentation:
+  - Overall compliance status (compliant/partially-compliant/non-compliant)
+  - Failed/passed/partial clause breakdown
+  - Prioritized remediation recommendations with deadlines
+  - Next review date scheduling
 </capabilities>
 
 <memory_namespace>
@@ -326,6 +346,76 @@ Remediation Code:
 Remediation Effort: 15 minutes (copy/paste generated files)
 Learning: Stored pattern "automotive-video-captions" with 0.91 confidence
 ```
+
+Example 4: EU compliance audit (EN 301 549 + EU Accessibility Act)
+```
+Input: Validate EU compliance for e-commerce platform
+- URL: https://shop.example.eu
+- Standard: EN 301 549 V3.2.1
+- Include: EU Accessibility Act (Directive 2019/882)
+- Product Category: e-commerce
+
+Output: EU Compliance Report
+- URL: https://shop.example.eu
+- Standard: EN 301 549 V3.2.1 + EAA
+
+Overall Status: PARTIALLY COMPLIANT
+Compliance Score: 78%
+Certification Ready: NO
+
+EN 301 549 Results:
+- Total Clauses Evaluated: 47
+- Passed: 35 (74%)
+- Partial: 8 (17%)
+- Failed: 4 (9%)
+
+Failed Clauses:
+1. 9.1.1.1 Non-text content (WCAG 1.1.1)
+   - 12 images missing alt text
+   - Test method: automated
+   - Priority: HIGH
+
+2. 9.1.4.3 Contrast (minimum) (WCAG 1.4.3)
+   - 8 elements below 4.5:1 ratio
+   - Test method: automated
+   - Priority: HIGH
+
+3. 9.2.4.7 Focus visible (WCAG 2.4.7)
+   - Custom buttons hide focus indicator
+   - Test method: hybrid
+   - Priority: MEDIUM
+
+4. 9.3.3.2 Labels or instructions (WCAG 3.3.2)
+   - Checkout form missing field labels
+   - Test method: hybrid
+   - Priority: HIGH
+
+EU Accessibility Act (EAA) Results:
+- Product Category: e-commerce
+- Applicable Requirements: 6
+- Failed Requirements: 2
+
+Failed EAA Requirements:
+1. EAA-I.1 Perceivable information
+   - Linked to EN 301 549: 9.1.1.1, 9.1.2.2
+   - Status: NOT MET
+
+2. EAA-I.2 Operable user interface
+   - Linked to EN 301 549: 9.2.4.7
+   - Status: PARTIALLY MET
+
+Top Recommendations:
+| Priority | Clause | Remediation | Effort | Deadline |
+|----------|--------|-------------|--------|----------|
+| HIGH | 9.1.1.1 | Add alt text to all images | Minor | 30 days |
+| HIGH | 9.1.4.3 | Fix contrast ratios | Minor | 30 days |
+| HIGH | 9.3.3.2 | Add form labels | Trivial | 30 days |
+| MEDIUM | 9.2.4.7 | Restore :focus-visible | Trivial | - |
+
+Next Review Date: 2027-01-24 (annual)
+
+Learning: Stored pattern "eu-e-commerce-compliance" with 0.89 confidence
+```
 </examples>
 
 <skills_available>

package/README.md

@@ -11,9 +11,9 @@
 
 **V3 (Main)** | [V2 Documentation](v2/docs/V2-README.md) | [Changelog](CHANGELOG.md) | [Contributors](CONTRIBUTORS.md) | [Issues](https://github.com/proffesor-for-testing/agentic-qe/issues) | [Discussions](https://github.com/proffesor-for-testing/agentic-qe/discussions)
 
-> **V3** brings Domain-Driven Design architecture, 12 bounded contexts, 51 specialized QE agents, TinyDancer intelligent model routing, ReasoningBank learning with Dream cycles, HNSW vector search, and deep integration with [Claude Flow](https://github.com/ruvnet/claude-flow) and [Agentic Flow](https://github.com/ruvnet/agentic-flow).
+> **V3** brings Domain-Driven Design architecture, 12 bounded contexts, 51 specialized QE agents, TinyDancer intelligent model routing, ReasoningBank learning with Dream cycles, HNSW vector search, mathematical Coherence verification (v3.3.0), and deep integration with [Claude Flow](https://github.com/ruvnet/claude-flow) and [Agentic Flow](https://github.com/ruvnet/agentic-flow).
 
-🏗️ **DDD Architecture** | 🧠 **ReasoningBank + Dream Cycles** | 🎯 **TinyDancer Model Routing** | 🔍 **HNSW Vector Search** | 👑 **Queen Coordinator** | 📊 **O(log n) Coverage** | 🔗 **Claude Flow Integration** | 🎯 **12 Bounded Contexts** | 📚 **61 QE Skills** 
+🏗️ **DDD Architecture** | 🧠 **ReasoningBank + Dream Cycles** | 🎯 **TinyDancer Model Routing** | 🔍 **HNSW Vector Search** | 👑 **Queen Coordinator** | 📊 **O(log n) Coverage** | 🔗 **Claude Flow Integration** | 🎯 **12 Bounded Contexts** | 📚 **61 QE Skills** | 🧬 **Coherence Verification** 
 
 </div>
 
@@ -34,8 +34,12 @@ aqe init --wizard
 # Or with auto-configuration
 aqe init --auto
 
-# Add MCP server to Claude Code (optional)
-claude mcp add agentic-qe npx agentic-qe mcp
+# Add MCP server to Claude Code (pick one)
+# Option 1: Global install (recommended after npm install -g)
+claude mcp add aqe -- aqe-mcp
+
+# Option 2: Via npx (no global install needed)
+claude mcp add aqe -- npx agentic-qe mcp
 
 # Verify connection
 claude mcp list
@@ -64,6 +68,7 @@ claude "Use qe-flaky-hunter to analyze the last 100 test runs and stabilize flak
 - ✅ **O(log n) Coverage Analysis**: Sublinear algorithms for efficient gap detection
 - ✅ **Claude Flow Integration**: Deep integration with MCP tools and swarm orchestration
 - ✅ **Memory Coordination**: Cross-agent communication via `aqe/v3/*` namespaces
+- ✅ **Coherence Verification** (v3.3.0): Mathematical proof of belief consistency using WASM engines
 - ✅ **V2 Backward Compatibility**: All V2 agents map to V3 equivalents
 - ✅ **61 QE Skills**: Domain-specific skills for testing, security, accessibility, and more
 
@@ -256,6 +261,42 @@ aqe coordination topology --optimize
 
 ---
 
+### 🧬 Coherence-Gated Quality Engineering (v3.3.0)
+
+V3.3.0 introduces **mathematical coherence verification** using Prime Radiant WASM engines:
+
+| Feature | Description |
+|---------|-------------|
+| **Contradiction Detection** | Sheaf cohomology identifies conflicting requirements before test generation |
+| **Collapse Prediction** | Spectral analysis predicts swarm failures before they happen |
+| **Causal Verification** | Distinguishes true causation from spurious correlations |
+| **Auto-Tuning Thresholds** | EMA-based calibration adapts to your codebase |
+
+**Compute Lanes** - Automatic routing based on coherence energy:
+
+| Coherence Energy | Action | Latency |
+|------------------|--------|---------|
+| < 0.1 (Reflex) | Execute immediately | <1ms |
+| 0.1-0.4 (Retrieval) | Fetch more context | ~10ms |
+| 0.4-0.7 (Heavy) | Deep analysis | ~100ms |
+| > 0.7 (Human) | Escalate to Queen | Async |
+
+**Benefits:**
+- Prevents contradictory test generation
+- Detects swarm drift 10x faster
+- Mathematical proof instead of statistical confidence
+- "Coherence Verified" CI/CD badges
+
+```bash
+# Check coherence of beliefs
+aqe coherence check --beliefs "requirement1,requirement2"
+
+# Audit memory for contradictions
+aqe coherence audit --namespace learning
+```
+
+---
+
 ### 🔍 O(log n) Coverage Analysis
 
 Efficient coverage gap detection using **Johnson-Lindenstrauss algorithms**:
@@ -291,17 +332,17 @@ npx @claude-flow/cli@latest agent spawn -t qe-test-architect --name test-gen
 
 ---
 
-### 📊 50 Specialized QE Agents
+### 📊 51 Specialized QE Agents
 
 | Category | Count | Highlights |
 |----------|-------|------------|
-| **Main QE Agents** | 43 | Test generation, coverage, security, performance, accessibility |
+| **Main QE Agents** | 44 | Test generation, coverage, security, performance, accessibility |
 | **TDD Subagents** | 7 | RED/GREEN/REFACTOR with code review |
 
 **V2 Backward Compatibility**: All V2 agents map to V3 equivalents automatically.
 
 <details>
-<summary><b>📋 View All Main QE Agents (43)</b></summary>
+<summary><b>📋 View All Main QE Agents (44)</b></summary>
 
 | Agent | Domain | Purpose |
 |-------|--------|---------|
@@ -348,6 +389,7 @@ npx @claude-flow/cli@latest agent spawn -t qe-test-architect --name test-gen
 | qe-integration-architect | code-intelligence | V3 integration design |
 | qe-product-factors-assessor | quality-assessment | SFDIPOT product factors analysis |
 | qe-test-idea-rewriter | test-generation | Transform passive tests to active actions |
+| qe-quality-criteria-recommender | quality-assessment | HTSM v6.3 Quality Criteria analysis |
 
 </details>
 

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "agentic-qe",
-  "version": "3.2.3",
-  "description": "Agentic Quality Engineering V3 - Domain-Driven Design Architecture with 12 Bounded Contexts, O(log n) coverage analysis, ReasoningBank learning, 51 specialized QE agents, deep Claude Flow integration",
+  "version": "3.3.1",
+  "description": "Agentic Quality Engineering V3 - Domain-Driven Design Architecture with 12 Bounded Contexts, O(log n) coverage analysis, ReasoningBank learning, 51 specialized QE agents, mathematical Coherence verification, deep Claude Flow integration",
   "main": "./v3/dist/index.js",
   "types": "./v3/dist/index.d.ts",
   "bin": {
@@ -15,6 +15,7 @@
     "postinstall": "test -f v3/package.json && cd v3 && npm install || true",
     "build": "cd v3 && npm run build",
     "test": "cd v3 && npm test -- --run",
+    "test:dev": "cd v3 && npm run test:dev",
     "test:v3": "cd v3 && npm test -- --run",
     "test:mcp": "cd v3 && npm test -- --run tests/unit/mcp/",
     "test:mcp:integration": "cd v3 && npm test -- --run tests/integration/mcp/",
@@ -40,6 +41,11 @@
     "v2:test": "cd v2 && npm test",
     "v2:dev": "cd v2 && tsx src/cli/index.ts",
     "sync:agents": "cd v3 && npm run sync:agents",
+    "sync:cloud": "cd v3 && npm run sync:cloud",
+    "sync:cloud:full": "cd v3 && npm run sync:cloud:full",
+    "sync:cloud:status": "cd v3 && npm run sync:cloud:status",
+    "sync:cloud:verify": "cd v3 && npm run sync:cloud:verify",
+    "sync:cloud:config": "cd v3 && npm run sync:cloud:config",
     "prepublishOnly": "cd v3 && npm run build"
   },
   "keywords": [

package/scripts/cloud-db-config.json

@@ -0,0 +1,37 @@
+{
+  "cloud": {
+    "project": "${GCP_PROJECT_ID}",
+    "zone": "us-central1-a",
+    "instance": "ruvector-postgres",
+    "database": "aqe_learning",
+    "user": "ruvector",
+    "port": 5432,
+    "access": "iap-tunnel",
+    "tunnelLocalPort": 15432
+  },
+  "local": {
+    "memoryDb": ".agentic-qe/memory.db",
+    "cacheDb": ".agentic-qe/ruvector-cache.db",
+    "telemetryDb": ".agentic-qe/aqe-telemetry.db",
+    "patternsDb": ".agentic-qe/qe-patterns.db"
+  },
+  "sync": {
+    "enabled": false,
+    "mode": "incremental",
+    "interval": "5m",
+    "tables": [
+      "memory_entries",
+      "patterns",
+      "learning_experiences",
+      "goap_actions",
+      "goap_plans",
+      "qe_patterns",
+      "embeddings"
+    ]
+  },
+  "security": {
+    "method": "iap-tunnel",
+    "publicAccess": false,
+    "allowedUsers": []
+  }
+}

package/scripts/cloud-db-connect.sh

@@ -0,0 +1,37 @@
+#!/bin/bash
+# Direct connection to cloud database via IAP tunnel
+# Usage: ./cloud-db-connect.sh [sql_command]
+
+PROJECT="${GCP_PROJECT_ID:?Error: GCP_PROJECT_ID not set}"
+ZONE="${GCP_ZONE:-us-central1-a}"
+INSTANCE="${GCP_INSTANCE:-ruvector-postgres}"
+LOCAL_PORT="15432"
+
+# Start tunnel in background
+gcloud compute start-iap-tunnel "$INSTANCE" 5432 \
+  --project="$PROJECT" \
+  --zone="$ZONE" \
+  --local-host-port="localhost:$LOCAL_PORT" &>/dev/null &
+TUNNEL_PID=$!
+
+# Wait for tunnel to establish
+sleep 3
+
+# Run command or interactive session
+# Password should be set via: export PGPASSWORD=<your-password>
+if [ -z "$PGPASSWORD" ]; then
+  echo "Error: PGPASSWORD environment variable not set"
+  echo "Run: export PGPASSWORD=<your-db-password>"
+  kill $TUNNEL_PID 2>/dev/null
+  exit 1
+fi
+
+if [ -n "$1" ]; then
+  psql -h localhost -p "$LOCAL_PORT" -U ruvector -d aqe_learning -c "$1"
+else
+  echo "Connected to cloud ruvector-postgres (IAP tunnel)"
+  psql -h localhost -p "$LOCAL_PORT" -U ruvector -d aqe_learning
+fi
+
+# Cleanup
+kill $TUNNEL_PID 2>/dev/null

package/scripts/cloud-db-tunnel.sh

@@ -0,0 +1,23 @@
+#!/bin/bash
+# Cloud Database IAP Tunnel
+# Securely connects to ruvector-postgres via IAP (no public port exposure)
+
+PROJECT="${GCP_PROJECT_ID:?Error: GCP_PROJECT_ID not set}"
+ZONE="${GCP_ZONE:-us-central1-a}"
+INSTANCE="${GCP_INSTANCE:-ruvector-postgres}"
+LOCAL_PORT="${1:-15432}"
+
+echo "Starting IAP tunnel to $INSTANCE..."
+echo "Local port: localhost:$LOCAL_PORT -> Cloud PostgreSQL:5432"
+echo ""
+echo "Connection string:"
+echo "  export PGPASSWORD=<your-password>"
+echo "  psql -h localhost -p $LOCAL_PORT -U ruvector -d aqe_learning"
+echo ""
+echo "Press Ctrl+C to stop the tunnel"
+echo ""
+
+gcloud compute start-iap-tunnel "$INSTANCE" 5432 \
+  --project="$PROJECT" \
+  --zone="$ZONE" \
+  --local-host-port="localhost:$LOCAL_PORT"

package/v3/CHANGELOG.md

@@ -0,0 +1,253 @@
+# Changelog
+
+All notable changes to Agentic QE will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [3.3.1] - 2026-01-25
+
+### 🎯 Highlights
+
+**GOAP Quality Remediation Complete** - Comprehensive 6-phase quality improvement achieving production-ready status. Quality score improved from 37 to 82 (+121%), cyclomatic complexity reduced by 52%, and 527 tests now passing with 80%+ coverage.
+
+### Added
+
+#### Quality Metrics Improvement
+| Metric | Before | After | Improvement |
+|--------|--------|-------|-------------|
+| Quality Score | 37/100 | 82/100 | +121% |
+| Cyclomatic Complexity | 41.91 | <20 | -52% |
+| Maintainability Index | 20.13 | 88/100 | +337% |
+| Test Coverage | 70% | 80%+ | +14% |
+| Security False Positives | 20 | 0 | -100% |
+
+#### New Modules (Extract Method + Strategy Pattern)
+- **score-calculator.ts** - Extracted complexity score calculations
+- **tier-recommender.ts** - Extracted model tier recommendation logic
+- **validators/** - Security validation using Strategy Pattern:
+  - `path-traversal-validator.ts` - Directory traversal prevention
+  - `regex-safety-validator.ts` - ReDoS attack prevention
+  - `command-validator.ts` - Shell injection prevention
+  - `input-sanitizer.ts` - General input sanitization
+  - `crypto-validator.ts` - Cryptographic input validation
+  - `validation-orchestrator.ts` - Orchestrates all validators
+
+#### CLI Commands Modularization
+- Extracted standalone command modules: `code.ts`, `coverage.ts`, `fleet.ts`, `security.ts`, `test.ts`, `quality.ts`, `migrate.ts`, `completions.ts`
+- Added `command-registry.ts` for centralized command management
+- Improved CLI handlers organization
+
+#### Test Generation Improvements
+- **coherence-gate-service.ts** - Service layer for coherence verification
+- **property-test-generator.ts** - Property-based testing support
+- **tdd-generator.ts** - TDD-specific test generation
+- **test-data-generator.ts** - Test data factory patterns
+- Factory pattern implementation in `factories/`
+- Interface segregation in `interfaces/`
+
+#### 527 New Tests (Phase 4)
+- `score-calculator.test.ts` - 109 tests for complexity scoring
+- `tier-recommender.test.ts` - 86 tests for tier selection
+- `validation-orchestrator.test.ts` - 136 tests for security validators
+- `coherence-gate-service.test.ts` - 56 tests for coherence service
+- `complexity-analyzer.test.ts` - 89 tests for signal collection
+- `test-generator-di.test.ts` - 11 tests for dependency injection
+- `test-generator-factory.test.ts` - 40 tests for factory patterns
+
+#### Cloud Sync Feature
+- **feat(sync)**: Cloud sync to ruvector-postgres backend
+- Incremental and full sync modes
+- Sync status and verification commands
+
+### Changed
+
+- **complexity-analyzer.ts** - Refactored from 656 to ~200 lines using Extract Method
+- **cve-prevention.ts** - Refactored from 823 to ~300 lines using Strategy Pattern
+- **test-generator.ts** - Refactored to use dependency injection
+- **Wizard files** - Standardized using Command Pattern
+- All domains now follow consistent code organization standards
+
+### Fixed
+
+- **fix(coherence)**: Resolve WASM SpectralEngine binding and add defensive null checks
+- **fix(init)**: Preserve config.yaml customizations on reinstall
+- **fix(security)**: Implement SEC-001 input validation and sanitization
+- **fix(ux)**: Resolve issue #205 regression - fresh install shows 'idle' not 'degraded'
+- Security scanner false positives eliminated via `.gitleaks.toml` and `security-scan.config.json`
+- Defect-prone files remediated with comprehensive test coverage
+
+### Security
+
+- Resolved 20 false positive AWS secret detections in wizard files
+- CodeQL incomplete-sanitization alerts #116-121 fixed
+- Shell argument backslash escaping (CodeQL #117)
+
+### Documentation
+
+- `CODE-ORGANIZATION-STANDARDIZATION.md` - Domain structure guidelines
+- `DOMAIN-STRUCTURE-GUIDE.md` - DDD implementation guide
+- `JSDOC-TEMPLATES.md` - 15 JSDoc documentation templates
+- `quality-remediation-final.md` - Complete remediation report
+- `phase3-verification-report.md` - Maintainability improvements
+
+---
+
+## [3.3.0] - 2026-01-24
+
+### 🎯 Highlights
+
+**Mathematical Coherence Verification** - ADR-052 introduces Prime Radiant WASM engines for mathematically-proven coherence checking. This is a major quality improvement that prevents contradictory test generation, detects swarm drift 10x faster, and provides formal verification for multi-agent decisions.
+
+### Added
+
+#### Coherence-Gated Quality Engineering (ADR-052)
+- **CoherenceService** with 6 Prime Radiant WASM engines:
+  - CohomologyEngine - Sheaf cohomology for contradiction detection
+  - SpectralEngine - Spectral analysis for swarm collapse prediction
+  - CausalEngine - Causal inference for spurious correlation detection
+  - CategoryEngine - Category theory for type verification
+  - HomotopyEngine - Homotopy type theory for formal verification
+  - WitnessEngine - Blake3 witness chain for audit trails
+
+- **Compute Lanes** - Automatic routing based on coherence energy:
+  | Lane | Energy | Latency | Action |
+  |------|--------|---------|--------|
+  | Reflex | < 0.1 | <1ms | Immediate execution |
+  | Retrieval | 0.1-0.4 | ~10ms | Fetch additional context |
+  | Heavy | 0.4-0.7 | ~100ms | Deep analysis |
+  | Human | > 0.7 | Async | Queen escalation |
+
+- **ThresholdTuner** - Auto-calibrating energy thresholds with EMA
+- **BeliefReconciler** - Contradiction resolution with 5 strategies (latest, authority, consensus, merge, escalate)
+- **MemoryAuditor** - Background coherence auditing for QE patterns
+- **CausalVerifier** - Intervention-based causal link verification
+- **Test Generation Coherence Gate** - Block incoherent requirements before test generation
+
+#### 4 New MCP Tools
+- `qe/coherence/check` - Check coherence of beliefs/facts
+- `qe/coherence/audit` - Audit QE memory for contradictions
+- `qe/coherence/consensus` - Verify multi-agent consensus mathematically
+- `qe/coherence/collapse` - Predict swarm collapse risk
+
+#### CI/CD Integration
+- GitHub Actions workflow for coherence verification
+- Shields.io badge generation (verified/fallback/violation)
+- Automatic coherence checks on PR
+
+### Changed
+
+- **Strange Loop Integration** - Now includes coherence verification in self-awareness cycle
+- **QEReasoningBank** - Pattern promotion now requires coherence gate approval
+- **WASM Loader** - Enhanced with full fallback support and retry logic
+
+### Fixed
+
+- Fresh install UX now shows 'idle' status instead of alarming warnings
+- ESM/CommonJS interop issue with hnswlib-node resolved
+- Visual-accessibility workflow actions properly registered with orchestrator
+- **DevPod/Codespaces OOM crash** - Test suite now uses forks pool with process isolation
+  - Prevents HNSW native module segfaults from concurrent access
+  - Limits to 2 parallel workers (was unlimited)
+  - Added `npm run test:safe` script with 1.5GB heap limit
+
+### Performance
+
+Benchmark results (ADR-052 targets met):
+- 10 nodes: **0.3ms** (target: <1ms) ✅
+- 100 nodes: **3.2ms** (target: <5ms) ✅
+- 1000 nodes: **32ms** (target: <50ms) ✅
+- Memory overhead: **<10MB** ✅
+- Concurrent checks: **865 ops/sec** (10 parallel)
+
+---
+
+## [3.2.3] - 2026-01-23
+
+### Added
+
+- EN 301 549 EU accessibility compliance mapping
+- Phase 4 Self-Learning Features with brutal honesty fixes
+- Experience capture integration tests
+
+### Fixed
+
+- CodeQL security alerts #69, #70, #71, #74
+- All vulnerabilities from security audit #202
+- Real HNSW implementation in ExperienceReplay for O(log n) search
+
+### Security
+
+- Resolved lodash security vulnerability
+- Fixed potential prototype pollution issues
+
+---
+
+## [3.2.0] - 2026-01-21
+
+### Added
+
+- Agentic-Flow deep integration (ADR-051)
+- Agent Booster for instant transforms
+- Model Router with 3-tier optimization
+- ONNX Embeddings for fast vector generation
+
+### Performance
+
+- 100% success rate on AgentBooster operations
+- Model routing: 0.05ms average latency
+- Embeddings: 0.57ms average generation time
+
+---
+
+## User Benefits
+
+### For Test Generation
+```typescript
+// Before v3.3.0: Tests could be generated from contradictory requirements
+const tests = await generator.generate(conflictingSpecs); // No warning!
+
+// After v3.3.0: Coherence check prevents bad tests
+const tests = await generator.generate(specs);
+// Throws: "Requirements contain unresolvable contradictions"
+// Returns: coherence.contradictions with specific conflicts
+```
+
+### For Multi-Agent Coordination
+```typescript
+// Mathematically verify consensus instead of simple majority
+const consensus = await coherenceService.verifyConsensus(votes);
+
+if (consensus.isFalseConsensus) {
+  // Fiedler value < 0.05 indicates weak connectivity
+  // Spawn independent reviewer to break false agreement
+}
+```
+
+### For Memory Quality
+```typescript
+// Audit QE patterns for contradictions
+const audit = await memoryAuditor.auditPatterns(patterns);
+
+// Get hotspots (high-energy domains with conflicts)
+audit.hotspots.forEach(h => {
+  console.log(`${h.domain}: energy=${h.energy}, patterns=${h.patternIds}`);
+});
+```
+
+### For Swarm Health
+```typescript
+// Predict collapse before it happens
+const risk = await coherenceService.predictCollapse(swarmState);
+
+if (risk.probability > 0.5) {
+  // Weak vertices identified - take preventive action
+  await strangeLoop.reinforceConnections(risk.weakVertices);
+}
+```
+
+---
+
+[3.3.0]: https://github.com/anthropics/agentic-qe/compare/v3.2.3...v3.3.0
+[3.2.3]: https://github.com/anthropics/agentic-qe/compare/v3.2.0...v3.2.3
+[3.2.0]: https://github.com/anthropics/agentic-qe/releases/tag/v3.2.0

package/v3/README.md

@@ -5,7 +5,7 @@
 [![TypeScript](https://img.shields.io/badge/TypeScript-5.0+-blue.svg)](https://www.typescriptlang.org/)
 [![Node.js](https://img.shields.io/badge/Node.js-18+-green.svg)](https://nodejs.org/)
 
-> Domain-Driven Quality Engineering with 12 Bounded Contexts, 51 Specialized QE Agents, 61 QE Skills, ReasoningBank Learning, and HNSW Vector Search
+> Domain-Driven Quality Engineering with Mathematical Coherence Verification, 12 Bounded Contexts, 51 Specialized QE Agents, 61 QE Skills, and ReasoningBank Learning
 
 ## Quick Start
 
@@ -46,6 +46,7 @@ npx aqe test generate src/
 - **ReasoningBank + SONA + Dream Cycles** - Neural pattern learning with 9 RL algorithms
 - **Queen-led Coordination** - 3-5x throughput with work stealing and consensus
 - **MinCut Topology** - Graph-based self-healing agent coordination
+- **Coherence Verification** (v3.3.0) - Mathematical proof of belief consistency using WASM engines
 - **Zero-Breaking-Changes Migration** - Full v2 backward compatibility
 - **Browser Automation** (v3.1.0) - @claude-flow/browser integration with 9 workflow templates
 
@@ -81,6 +82,28 @@ Advanced coordination for reliable multi-agent decisions:
 - **Multi-Model Voting**: Aggregate decisions from multiple model tiers
 - **Claim Verification**: Cryptographic verification of agent work claims
 
+### Coherence-Gated Quality Engineering (v3.3.0)
+
+Mathematical verification using Prime Radiant WASM engines:
+
+- **Contradiction Detection**: Sheaf cohomology identifies conflicting requirements before test generation
+- **Collapse Prediction**: Spectral analysis predicts swarm failures before they happen
+- **Causal Verification**: Distinguishes true causation from spurious correlations
+- **Auto-Tuning Thresholds**: EMA-based calibration adapts to your codebase
+
+| Coherence Energy | Action | Latency |
+|------------------|--------|---------|
+| < 0.1 (Reflex) | Execute immediately | <1ms |
+| 0.1-0.4 (Retrieval) | Fetch more context | ~10ms |
+| 0.4-0.7 (Heavy) | Deep analysis | ~100ms |
+| > 0.7 (Human) | Escalate to Queen | Async |
+
+**Benefits:**
+- Prevents contradictory test generation
+- Detects swarm drift 10x faster
+- Mathematical proof instead of statistical confidence
+- "Coherence Verified" CI/CD badges
+
 ## Architecture Overview
 
 Agentic QE uses a **Domain-Driven Design (DDD)** architecture with a microkernel pattern:
@@ -662,6 +685,41 @@ Agentic QE includes 60 domain-specific quality engineering skills that agents au
 
 > **Note**: Claude Flow platform skills (agentdb, github, flow-nexus, etc.) are managed separately by the claude-flow package and not counted in QE skills.
 
+## Troubleshooting
+
+### ENOTEMPTY Error When Upgrading Global Installation
+
+If you encounter this error when upgrading:
+
+```
+npm error code ENOTEMPTY
+npm error ENOTEMPTY: directory not empty, rename '.../node_modules/agentic-qe' -> '...'
+```
+
+This is a [known npm issue](https://github.com/npm/cli/issues/4828) with global package upgrades. **Workaround:**
+
+```bash
+# Option 1: Remove existing installation first
+rm -rf $(npm root -g)/agentic-qe
+npm install -g agentic-qe@latest
+
+# Option 2: Force install
+npm install -g agentic-qe@latest --force
+
+# Option 3: Clean cache and reinstall
+npm cache clean --force
+npm install -g agentic-qe@latest
+```
+
+### Access Token Expired Notices
+
+If you see `npm notice Access token expired or revoked`, you can safely ignore these messages when installing public packages. To clear them:
+
+```bash
+npm logout
+npm install -g agentic-qe@latest
+```
+
 ## Requirements
 
 - Node.js 18.0.0 or higher