← ClaudeAtlas

craft-testinglisted

Craftsman standard for automated testing: strategy, unit/integration/e2e selection, refactor-proof design, flaky tests, mocking boundaries, deterministic data, and merge-gate policy. Use WHENEVER work touches tests: writing/reviewing tests, strategy, "add tests", "why is this flaky", "what should I test", "tests pass but prod breaks", Testing Library / Playwright / Vitest / Jest / Pytest, mocks, fixtures, or adequacy. Trigger on "add tests", "write a test", "is this tested enough", or "make the tests reliable". Owns which suites gate merge and what "green" means (incl. e2e); CI pipeline mechanism → craft-infra (missing e2e suite → TEST; e2e not in CI → INFRA). Security correctness → craft-security; visual live audit → craft-ux. Whole-project readiness → craft-audit. Existing findings ("fix TEST-004") → craft-fix.
atifgul99/craftsman-marketplace · ★ 1 · Testing & QA · score 65
Install: claude install-skill atifgul99/craftsman-marketplace
# Testing Craft This skill encodes one engineer's standard for testing software so the suite is actually trusted — fast, deterministic, and catching real regressions rather than decorating the coverage badge. The **method and opinions** live here; the **project specifics** (test runner, framework, what's already covered) live in the target repo — always discover them, never assume or hardcode. The persona this serves usually arrives at one of two extremes: **no tests at all**, or a pile of **AI-generated tests that don't test anything** — they assert that a mock was called, pin implementation details, or cover trivial getters to hit a number while the payment path has zero coverage. Both feel like "we have testing." Neither catches the bug that takes the app down. The job is to move them to a small set of tests they can *trust*, on the paths that actually matter. ## Operating principle — discover before you build Different repos already have different pieces in place. Before adding anything, map what exists so you extend rather than duplicate or fight it: - `package.json` / lockfile → test runner (`vitest`, `jest`, `playwright`, `@testing-library/*`, `pytest`), `test`/`test:e2e`/`coverage` scripts, and whether tests even run. Coverage tooling or a configured percentage is context, not proof that the tests are adequate. - Test config (`vitest.config.*`, `jest.config.*`, `playwright.config.*`, `pytest.ini`, `conftest.py`) → environment, setup files, coverage threshol