| name | python-expert |
| description | Use for Python code changes, FastAPI/Django/Flask services, packaging, typing, async, pytest, data scripts, dependency hygiene, or Python performance fixes. |
| version | 1.0.0 |
| author | awesome-grok-build |
Python Expert
Ship idiomatic Python with tests, typing, and boring operational behavior.
Grok Build Mode
- Use Plan Mode for package layout changes, API changes, async rewrites, migrations, or dependency upgrades.
- Use subagents when scope is broad:
api: public interfaces, routes, schemas, CLI commands.
tests: pytest coverage and fixtures.
types: mypy/pyright/ruff/typing issues.
runtime: async, IO, DB queries, performance.
- Arena-style comparison: if Arena Mode is available, compare alternative implementations for risky refactors. Otherwise ask subagents for independent approaches and choose the simplest verified path.
Workflow
- Detect project tooling from
pyproject.toml, setup.cfg, tox.ini, noxfile.py, requirements*.txt, and CI.
- Follow existing style and framework patterns.
- Prefer small functions, explicit errors, and typed boundaries.
- Add or update pytest tests for changed behavior.
- Run the narrowest relevant command:
pytest path/to/test.py -q
ruff check .
ruff format .
mypy . or pyright
- Summarize changed files, tests, and residual risk.
Python Rules
- Do not add dependencies without a plan.
- Keep sync/async boundaries explicit.
- Use context managers for files, network clients, DB sessions, and locks.
- Validate external inputs with the project's existing validation layer.
- Preserve public APIs unless a breaking change is approved.
- Prefer
pathlib, dataclasses/Pydantic where already used, and structured logging.
Example Prompts
Use python-expert. Add tests first, then fix this FastAPI bug. Run the narrowest pytest command.
Use python-expert with subagents for API, tests, types, and runtime. Plan before editing.