Agentic QE -- AI-Powered Quality Engineering
Comprehensive Quality Engineering plugin with 51 specialized agents across 12 DDD bounded contexts. Features AI-powered test generation, O(log n) coverage analysis, defect prediction, security compliance, contract testing, visual/accessibility testing, and chaos engineering.
Purpose
Automate all aspects of software quality: write tests, find coverage gaps, predict bugs, scan for vulnerabilities, and test system resilience. Think of it as a full QA team that never sleeps and learns from every bug it finds.
Bounded Contexts (12)
| Context |
Agents |
Description |
| test-generation |
12 |
Unit, integration, E2E, property, mutation, fuzz, API, performance, security, accessibility, contract, BDD |
| test-execution |
8 |
Runner, parallel executor, retry manager, result aggregator, flaky detector, timeout manager, resource allocator, reporter |
| coverage-analysis |
6 |
Collector, gap detector, priority ranker, hotspot analyzer, trend tracker, impact assessor |
| quality-assessment |
5 |
Gate evaluator, readiness assessor, risk calculator, metric aggregator, decision maker |
| defect-intelligence |
4 |
Predictor, root cause analyzer, pattern detector, regression tracker |
| requirements-validation |
3 |
BDD validator, testability analyzer, requirement tracer |
| code-intelligence |
5 |
Knowledge graph builder, semantic searcher, dependency analyzer, complexity assessor, pattern miner |
| security-compliance |
4 |
SAST scanner, DAST scanner, audit trail manager, compliance checker |
| contract-testing |
3 |
OpenAPI, GraphQL, gRPC validators |
| visual-accessibility |
3 |
Visual regression detector, WCAG checker, screenshot differ |
| chaos-resilience |
4 |
Chaos injector, load generator, resilience assessor, recovery validator |
| learning-optimization |
2 |
Cross-domain learner, pattern optimizer |
MCP Tools (16)
Test Generation
npx claude-flow@v3alpha mcp call aqe/generate-tests \
--targetPath ./src/utils/calculator.ts \
--testType unit --framework vitest
npx claude-flow@v3alpha mcp call aqe/tdd-cycle \
--requirement "Users can reset password via email" \
--targetPath ./src/auth --style london
Coverage Analysis
npx claude-flow@v3alpha mcp call aqe/analyze-coverage \
--targetPath ./src --threshold 80
npx claude-flow@v3alpha mcp call aqe/prioritize-gaps \
--targetPath ./src --riskWeighted true
Quality Gates
npx claude-flow@v3alpha mcp call aqe/evaluate-quality-gate \
--gates '[{"metric":"line_coverage","operator":">=","threshold":80}]'
npx claude-flow@v3alpha mcp call aqe/assess-readiness \
--criteria '[{"name":"coverage","required":true}]'
Security Compliance
npx claude-flow@v3alpha mcp call aqe/security-scan \
--targetPath ./src --scanType sast --compliance owasp-top-10
Chaos Engineering
# Always dry-run first
npx claude-flow@v3alpha mcp call aqe/chaos-inject \
--target payment-service --failureType network-latency \
--duration 30 --intensity 0.5 --dryRun true
Failure types: network-latency, network-partition, cpu-stress, memory-pressure, disk-failure, process-kill, dns-failure.
TDD Subagents (London-Style)
7 specialized subagents for red-green-refactor cycles:
- requirement-analyzer -- Parse requirements into testable specifications
- test-designer -- Design test structure and cases
- red-phase-executor -- Write failing tests
- green-phase-implementer -- Write minimal code to pass
- refactor-advisor -- Suggest refactoring improvements
- coverage-verifier -- Verify coverage meets targets
- cycle-coordinator -- Orchestrate the full TDD cycle
Max cycles: 10. Default framework: vitest.
Model Routing (TinyDancer)
| Tier |
Handler |
Categories |
Latency |
| Tier 1 |
agent-booster |
add-test-import, add-assertion, remove-console |
<1ms |
| Tier 2 |
haiku |
generate-unit-test, generate-mock, analyze-coverage-line |
~500ms |
| Tier 2+ |
sonnet |
generate-integration-test, predict-defect-simple |
~2s |
| Tier 3 |
opus |
generate-e2e-test, root-cause-analysis, chaos-experiment-design |
~5s |
Memory Namespaces
aqe/v3/test-patterns -- Test generation patterns (permanent, 384-dim)aqe/v3/coverage-data -- Coverage data (24h TTL)aqe/v3/defect-patterns -- Bug prediction patterns (permanent)aqe/v3/security-findings -- Security scan results (permanent)aqe/v3/chaos-experiments -- Chaos test results (7-day TTL)aqe/v3/learning-trajectories -- Cross-domain learning (permanent)
Safety
- Chaos operations default to dry-run mode
- All code runs in sandbox: 30s timeout, 512MB memory limit, no network
- Production targets are blocked from chaos testing
- Allowed commands: node, npm, npx, vitest, jest, pytest
- Blocked paths: /etc, /var, ~/.ssh, ~/.aws
Performance Targets
| Metric |
Target |
| Test generation latency |
<2s |
| Coverage analysis |
O(log n) via Johnson-Lindenstrauss projection |
| Quality gate evaluation |
<500ms |
| Security scan per KLOC |
<10s |
| MCP tool response |
<100ms |
1---2name: agentic-qe3description: Quality Engineering with 51 agents across 12 DDD bounded contexts, TDD, chaos engineering4---56# Agentic QE -- AI-Powered Quality Engineering78Comprehensive Quality Engineering plugin with 51 specialized agents across 12 DDD bounded contexts. Features AI-powered test generation, O(log n) coverage analysis, defect prediction, security compliance, contract testing, visual/accessibility testing, and chaos engineering.910## Purpose1112Automate all aspects of software quality: write tests, find coverage gaps, predict bugs, scan for vulnerabilities, and test system resilience. Think of it as a full QA team that never sleeps and learns from every bug it finds.1314## Bounded Contexts (12)1516| Context | Agents | Description |17|-------------------------|--------|------------------------------------------------|18| test-generation | 12 | Unit, integration, E2E, property, mutation, fuzz, API, performance, security, accessibility, contract, BDD |19| test-execution | 8 | Runner, parallel executor, retry manager, result aggregator, flaky detector, timeout manager, resource allocator, reporter |20| coverage-analysis | 6 | Collector, gap detector, priority ranker, hotspot analyzer, trend tracker, impact assessor |21| quality-assessment | 5 | Gate evaluator, readiness assessor, risk calculator, metric aggregator, decision maker |22| defect-intelligence | 4 | Predictor, root cause analyzer, pattern detector, regression tracker |23| requirements-validation | 3 | BDD validator, testability analyzer, requirement tracer |24| code-intelligence | 5 | Knowledge graph builder, semantic searcher, dependency analyzer, complexity assessor, pattern miner |25| security-compliance | 4 | SAST scanner, DAST scanner, audit trail manager, compliance checker |26| contract-testing | 3 | OpenAPI, GraphQL, gRPC validators |27| visual-accessibility | 3 | Visual regression detector, WCAG checker, screenshot differ |28| chaos-resilience | 4 | Chaos injector, load generator, resilience assessor, recovery validator |29| learning-optimization | 2 | Cross-domain learner, pattern optimizer |3031## MCP Tools (16)3233### Test Generation34```bash35npx claude-flow@v3alpha mcp call aqe/generate-tests \36 --targetPath ./src/utils/calculator.ts \37 --testType unit --framework vitest3839npx claude-flow@v3alpha mcp call aqe/tdd-cycle \40 --requirement "Users can reset password via email" \41 --targetPath ./src/auth --style london42```4344### Coverage Analysis45```bash46npx claude-flow@v3alpha mcp call aqe/analyze-coverage \47 --targetPath ./src --threshold 804849npx claude-flow@v3alpha mcp call aqe/prioritize-gaps \50 --targetPath ./src --riskWeighted true51```5253### Quality Gates54```bash55npx claude-flow@v3alpha mcp call aqe/evaluate-quality-gate \56 --gates '[{"metric":"line_coverage","operator":">=","threshold":80}]'5758npx claude-flow@v3alpha mcp call aqe/assess-readiness \59 --criteria '[{"name":"coverage","required":true}]'60```6162### Security Compliance63```bash64npx claude-flow@v3alpha mcp call aqe/security-scan \65 --targetPath ./src --scanType sast --compliance owasp-top-1066```6768### Chaos Engineering69```bash70# Always dry-run first71npx claude-flow@v3alpha mcp call aqe/chaos-inject \72 --target payment-service --failureType network-latency \73 --duration 30 --intensity 0.5 --dryRun true74```7576Failure types: network-latency, network-partition, cpu-stress, memory-pressure, disk-failure, process-kill, dns-failure.7778## TDD Subagents (London-Style)79807 specialized subagents for red-green-refactor cycles:811. **requirement-analyzer** -- Parse requirements into testable specifications822. **test-designer** -- Design test structure and cases833. **red-phase-executor** -- Write failing tests844. **green-phase-implementer** -- Write minimal code to pass855. **refactor-advisor** -- Suggest refactoring improvements866. **coverage-verifier** -- Verify coverage meets targets877. **cycle-coordinator** -- Orchestrate the full TDD cycle8889Max cycles: 10. Default framework: vitest.9091## Model Routing (TinyDancer)9293| Tier | Handler | Categories | Latency |94|------|---------|------------|---------|95| Tier 1 | agent-booster | add-test-import, add-assertion, remove-console | <1ms |96| Tier 2 | haiku | generate-unit-test, generate-mock, analyze-coverage-line | ~500ms |97| Tier 2+ | sonnet | generate-integration-test, predict-defect-simple | ~2s |98| Tier 3 | opus | generate-e2e-test, root-cause-analysis, chaos-experiment-design | ~5s |99100## Memory Namespaces101102- `aqe/v3/test-patterns` -- Test generation patterns (permanent, 384-dim)103- `aqe/v3/coverage-data` -- Coverage data (24h TTL)104- `aqe/v3/defect-patterns` -- Bug prediction patterns (permanent)105- `aqe/v3/security-findings` -- Security scan results (permanent)106- `aqe/v3/chaos-experiments` -- Chaos test results (7-day TTL)107- `aqe/v3/learning-trajectories` -- Cross-domain learning (permanent)108109## Safety110111- Chaos operations default to dry-run mode112- All code runs in sandbox: 30s timeout, 512MB memory limit, no network113- Production targets are blocked from chaos testing114- Allowed commands: node, npm, npx, vitest, jest, pytest115- Blocked paths: /etc, /var, ~/.ssh, ~/.aws116117## Performance Targets118119| Metric | Target |120|--------|--------|121| Test generation latency | <2s |122| Coverage analysis | O(log n) via Johnson-Lindenstrauss projection |123| Quality gate evaluation | <500ms |124| Security scan per KLOC | <10s |125| MCP tool response | <100ms |
jrennie99-glitch