role-aqa:test-strategy

Test Strategy

When to use

• Starting a new project — need to choose pyramid, trophy, diamond, or honeycomb model
• Coverage is high but bugs still ship — need mutation testing or exploratory approach
• Designing CI/CD quality gates and shift-left checkpoints
• Risk-scoring features to allocate test effort proportionally
• Planning safe production validation (canary, feature flags, shadow traffic)
• Creating a formal test plan document with entry/exit criteria
• Choosing between Page Object Model, Screenplay, or App Actions patterns

Core principles

1. Test behavior, not implementation — tests survive refactoring; if touching internals breaks them, they're wrong
2. Model before metrics — choose pyramid/trophy/diamond/honeycomb first, then set coverage targets accordingly
3. Coverage is vanity, mutation score is truth — 80% line coverage with 0% mutation detection is theater
4. Shift left relentlessly — every bug caught at compile time costs 1x; the same bug in production costs 100x
5. Test in production safely — canary + synthetic monitors + feature flags beats pretending staging is production

Reference Files

• references/distribution-models.md — pyramid, trophy, diamond, honeycomb explained with ratios and when to pick each
• references/shift-left-risk-prioritization.md — shift-left pipeline gates, risk matrix scoring, test impact analysis tools, CI quality gate YAML
• references/mutation-exploratory-testing.md — Stryker/PITest/mutmut setup, mutation operators, SBTM exploratory charters, DORA and quality metrics
• references/testing-in-production.md — canary deployments, feature flags, shadow traffic, dark launching, synthetic monitoring, AI-assisted test generation
• references/test-architecture-planning.md — POM, Screenplay, App Actions patterns with code; regression strategy; environment topology; test plan template; Given/When/Then acceptance criteria