Testing Router

Testing Router

Don't know which testing skill to use? Start here. Describe what you need and get routed to the right specialist.

How It Works

1. Describe your testing need in plain language
2. The router matches your intent to the best specialist skill
3. Follow the specialist's workflow

Quick Decision Tree

What are you trying to do?
│
├── Writing NEW code?
│   ├── Want tests first? → tdd-workflow
│   ├── Want quality checks as you write? → plankton-code-quality
│   └── Building evals for AI behavior? → eval-harness
│
├── Testing EXISTING code?
│   ├── Unit/integration tests needed? → tdd-workflow
│   ├── End-to-end browser tests? → e2e-testing / webapp-testing
│   ├── Visual appearance tests? → visual-regression
│   ├── Performance/load tests? → load-testing
│   └── Python project? → python-testing
│
├── Finding BUGS?
│   ├── Find and fix bugs? → qa
│   ├── Find and report only? → qa-only
│   └── Check if recent changes broke things? → ai-regression-testing
│
├── VERIFYING work?
│   ├── Multi-step verification checklist? → verification-loop
│   └── Prove changes work before shipping? → verification-before-completion
│
└── PLANNING tests?
    └── What should I test and how? → test-strategy

Skill Capabilities Summary

tdd-workflow

• Red/green/refactor cycle
• Coverage tracking and enforcement (80% minimum)
• Works with Vitest, Jest, pytest, Go testing
• Best for: new features, bug reproduction

e2e-testing

• Playwright test generation
• Page Object Model patterns
• CI/CD integration (GitHub Actions)
• Best for: critical user journeys, cross-browser testing

webapp-testing

• Playwright with local dev server integration
• Auto-starts dev server, waits for ready
• Visual snapshot testing built in
• Best for: testing local web apps during development

visual-regression

• Screenshot comparison with configurable thresholds
• Percy and Chromatic integration
• Dynamic content handling (masking, freezing)
• Best for: UI consistency, design system validation

load-testing

• k6 and Artillery script generation
• Scenario types: ramp-up, sustained, spike, stress
• Threshold-based pass/fail (p95, error rate)
• Best for: API performance, capacity planning

qa / qa-only

• Systematic bug hunting across the codebase
• qa: finds AND fixes bugs, commits the fixes
• qa-only: finds and reports bugs in a structured report
• Best for: pre-release quality gates

ai-regression-testing

• AI-powered analysis of test result diffs
• Detects behavioral regressions beyond simple pass/fail
• Compares test output semantics across runs
• Best for: catching subtle regressions after refactors

verification-loop

• Multi-step verification checklist
• TypeScript compilation, lint, test suite, E2E
• Console error checking, manual verification prompts
• Best for: final sign-off before marking work complete

verification-before-completion

• Evidence-based proof that changes work
• Screenshots, test output, build logs as artifacts
• Structured proof document
• Best for: async review, proving to stakeholders

eval-harness

• Build evaluation criteria before implementation
• Define pass/fail for AI behaviors
• Structured eval suites with scoring
• Best for: AI/ML feature development, prompt engineering

plankton-code-quality

• Real-time quality checks during code authoring
• Catches common issues: mutation, deep nesting, missing error handling
• Language-aware patterns
• Best for: maintaining quality standards as you write

python-testing

• pytest patterns: fixtures, parametrize, mocking
• conftest.py organization
• Coverage with pytest-cov
• Best for: Python projects needing comprehensive test suites

test-strategy

• Analyzes project structure and recommends test approach
• Coverage targets by module criticality
• Framework recommendations based on stack
• Best for: greenfield projects, testing overhauls

Combining Skills

Skills compose naturally. Common combinations: