| agent | engineer |
| allowed-tools | ["Read","Bash","Grep","Glob","Edit","Write","LS"] |
| context | fork |
| description | Idiomatic Python 3.12+ development. Use when writing Python code, CLI tools, scripts, or services. Emphasizes stdlib, type hints, fast pytest feedback, uv/ruff/pyright toolchain, and minimal dependencies. NOT for Go, Rust, TypeScript, or shell-only tasks. |
| name | writing-python |
| user-invocable | false |
Python Development
Use this only for Python 3.12+ code. Keep language-agnostic rules out of this skill.
Read First
Read principles.md before generating or changing Python code.
Core Rules
- Prefer existing project patterns over these defaults.
- Inspect the project Python target before using 3.12-only syntax.
- Prefer stdlib and existing dependencies before adding packages.
- Validate untyped boundary input before it spreads.
Comments and Docstrings
- Prefer clear names and small functions over explanatory comments.
- Add comments only for non-obvious constraints, invariants, side effects, tradeoffs, or public contracts.
- Keep comments short. Move longer rationale to docs, issue links, or design notes.
- Do not comment obvious code or restate the syntax.
- Keep tests readable without comments; add one only for unobvious fixtures, timing, concurrency, or regression context.
- Use docstrings for public or non-obvious modules, classes, functions, and methods. Do not restate the signature.
Conditional References
- patterns.md — read for module design, typing, async, config, file I/O, and error patterns.
- cli.md — read before writing or changing Python CLIs.
- testing.md — read before adding or reshaping Python tests; treat slow tests as feedback-loop defects.
- linting.md — read before changing ruff or pyright config, or the lint/format/type-check flow.
Verification
Run project-configured tests, lint, format, and type checks. Prefer focused pytest commands for the edit loop, then the broader project command before final output. Prefer pytest, ruff, and pyright when present.
If a tool is not configured, say so and run the closest available gate. If a check fails, diagnose the cause, make a targeted fix, and rerun the relevant check.
Python-Specific Failure Cases
- No clear project root: locate
pyproject.toml before editing or choosing commands.
- Unknown Python target: inspect
pyproject.toml, .python-version, CI, or lockfiles before using 3.12-only syntax.
- Type checker reports missing attributes: check imports, package exports, and runtime shape before loosening types.
Final Response
Include:
- changed files
- checks run and results
- checks skipped with reasons
- remaining risks or follow-ups
