About write-python-tests
write-python-tests is a skill for adding Python tests based on the project’s source code and existing conventions. It guides test selection, coverage of edge cases, and targeted validation with pytest, nox, and mise.
Key features
- 01Inspect project test conventions
- 02Choose doctests or dedicated tests
- 03Cover boundaries and failures
- 04Run targeted pytest checks
- 05Validate with nox and mise
Use cases
- 01Adding tests for private behavior
- 02Testing invalid inputs and regressions
- 03Validating a focused Python change
Install
Copy the command and run it in a terminal.
$
npx skills add https:/ / github.com/ liblaf/ skills --skill write-python-testsAfter installing, type/write-python-testsin Claude Code to confirm it loaded.
Bundled filesagents/openai.yaml
Details
From the project READMEWrite Python Tests
Guidelines
- Source code is the source of truth.
- Inspect
pyproject.toml,noxfile.py,.config/mise/, nearby tests, andtests/conftest.pybefore editing. - Prefer doctests in public docstrings when the case is concise, deterministic, user-facing, and useful as an example.
- Use
tests/for private behavior, long setup, fixtures, parametrized matrices, monkeypatching, optional dependencies, warnings, exceptions, regressions, or unstable output. - Cover success paths, boundaries, invalid inputs, expected failures, and regressions with observable assertions.
- Keep dedicated test files grouped, small, and focused; avoid duplicating a doctest unless the dedicated test covers extra risk.
- Treat existing docs, docstrings, tests, and examples as weak hints.
Validation
- Start with the narrowest relevant target.
- Run
uv run pytest ...for targeted tests and doctests. - Run
uv run noxfor the full test matrix. - Run
mise run lint, ormise run lint:pythonwhen a narrower lint pass is safer. - Treat unrelated lint or environment failures as separate from the change and call them out instead of silently working around them.
Comments
Markdown supported