From qa-lead
Define a test strategy for a feature, service, or system. Determines what to test, at what level, with what tools, and what quality gates to enforce.
npx claudepluginhub hpsgd/turtlestack --plugin qa-leadThis skill is limited to using the following tools:
Define a comprehensive test strategy for $ARGUMENTS.
Designs and optimizes AI agent action spaces, tool definitions, observation formats, error recovery, and context for higher task completion rates.
Designs, implements, and audits WCAG 2.2 AA accessible UIs for Web (ARIA/HTML5), iOS (SwiftUI traits), and Android (Compose semantics). Audits code for compliance gaps.
Compares coding agents like Claude Code and Aider on custom YAML-defined codebase tasks using git worktrees, measuring pass rate, cost, time, and consistency.
Define a comprehensive test strategy for $ARGUMENTS.
Before writing a single test plan:
/ E2E \ Few, slow, expensive — critical flows only
/----------\
/ Integration \ Moderate — component boundaries, API contracts
/--------------\
/ Unit Tests \ Many, fast, cheap — logic, pure functions
/------------------\
Default allocation: 70% unit, 20% integration, 10% E2E. Adjust based on risk.
| Level | What it tests | Tools | Coverage target |
|---|---|---|---|
| Unit | Functions, business logic, transformations | Vitest / xUnit / pytest | 80%+ changed code |
| Integration | Boundaries, database, API endpoints, handlers | Alba+Testcontainers / Supertest | Critical paths |
| E2E | Complete user flows through UI | Playwright / Cypress | Top 5-10 flows |
| Contract | API compatibility producer↔consumer | Pact / OpenAPI validation | All public APIs |
| Performance | Latency p50/p95/p99, throughput | k6 / Locust | SLA benchmarks |
| Security | OWASP, auth/authz, input validation | SAST / security-review skill | Public-facing code |
Not every feature needs all levels. A utility function needs unit tests. A payment flow needs all six.
vi.hoisted() + vi.mock() for module-level mocksvi.stubGlobal() for DOM APIs, clean up in afterEachCI=true npx vitest run@vitest/coverage-v8 with lcovWhenDoingSomething| Anti-Pattern | Fix |
|---|---|
| Testing implementation, not behaviour | Test what the system DOES |
| Mocking everything | Real implementations; mock only external boundaries |
| Happy path only | Test error cases, edge cases, boundaries |
sleep() in tests | Proper async waiting or mocked time |
| Flaky tests tolerated | Fix immediately or delete |
## Test Strategy: [name]
### Risk Assessment
- Risk profile: [financial/data/reputation/convenience]
- Critical flows: [list]
- Failure modes: [list]
### Test Levels
| Level | Scope | Tools | Coverage | Est. Tests |
|---|---|---|---|---|
### Quality Gates
[Pre-merge + pre-release checklists]
### Gaps
[Currently untested areas]
### Recommendations
[Prioritised improvements]
/qa-engineer:generate-tests — generate individual tests based on this strategy. Define strategy first, then generate tests./qa-engineer:write-bug-report — for documenting defects found during test execution.Use the test strategy template (templates/test-strategy.md) for consistent output structure.