← ClaudeAtlas

code-testslisted

Load BEFORE writing or running any tests. Carries the project's testing conventions - how tests are run, where they live, the mocking rules, and the mandatory post-test checks (coverage, formatting, type-check). Trigger on "write a test", "add coverage", "fix this failing test", "run the tests".
NAVNAV221/ai-native-codebase · ★ 1 · Testing & QA · score 72
Install: claude install-skill NAVNAV221/ai-native-codebase
# Code Tests Tests are how a change earns the word "done". This skill carries the conventions a new session would otherwise have to reverse-engineer from existing tests - which is how bad patterns spread. ## Running State the one command that runs the suite from the repo root, e.g.: ```bash make test # or: pytest, npm test, go test ./... ``` ## Placement & naming - Where do tests live relative to the code they cover? - What is the file-naming convention? - Call out any surprises (tests that live somewhere non-obvious). ## Rules that matter here Replace with your project's real rules. Common ones worth stating explicitly: - Prefer real data objects to broad mocks for typed models. - Patch the symbol where it is *used*, not where it is defined. - Assert on behaviour and return values, not on log output. ## After the test passes 1. Check coverage on the changed files. 2. Run the formatter. 3. Run the type-checker on the changed files. A test that passes but leaves the formatter or type-checker red is not done.