Testing
Overview
Six leaves cover the automated-testing craft: the portfolio and its gates, the design of one test, the red-green-refactor cycle, adequacy measures beyond coverage, the flake triage loop, and browser-level end-to-end checks. Strategy chooses what to test; the rest do the testing.
When to use
- Choosing levels, coverage policy and CI test gates for a project.
- Designing unit, integration or contract tests: doubles, fixtures, data.
- Working test-first, or writing the failing reproduction test for a bug fix.
- Judging suite strength with mutation, property or fuzz testing.
- Diagnosing and fixing a nondeterministic test.
- Writing Playwright/Cypress-style browser tests for a critical journey.
When not to use
- Reproducing or fixing the product bug the test exposed — outside this skill.
- Wiring the CI pipeline itself, its caching or its gate order — outside this skill.
- Accessibility conformance audits — outside this skill.
Precedence
The deepest loaded instruction governs its uniquely-claimed scope; every
scope:string is claimed by exactly one file in this tree; no leaf may widen a prohibition declared by an ancestor — leaves narrow, never override upward.
Subskill index
| Leaf | Task | Trigger |
|---|---|---|
subskills/test-strategy.md |
Test portfolio and CI gates | "what should we test" |
subskills/writing-tests.md |
Design one unit or integration test | "write tests for this" |
subskills/tdd.md |
Red-green-refactor | "do it test-first" |
subskills/test-adequacy.md |
Property, mutation, fuzz testing | "are my tests good enough" |
subskills/flaky-tests.md |
Nondeterministic tests | "this test fails randomly" |
subskills/browser-e2e-testing.md |
Browser and end-to-end tests | "write a Playwright test" |
See also
subskills/test-strategy.md— pick the portfolio.subskills/writing-tests.md— build one test right.subskills/tdd.md— the test-first loop.subskills/test-adequacy.md— prove the suite catches mutants.subskills/flaky-tests.md— kill the flake.subskills/browser-e2e-testing.md— the thin top layer.