python-engineering

This skill improves how Python work is planned, implemented, and reviewed. Use it for scripts, packages, CLIs, FastAPI services, notebooks, automation, and data-heavy Python tasks.

Quick Start

1. Classify the request: new code, debugging, refactoring, optimization, tests, architecture, notebook cleanup, or review.
2. Inspect local context first: Python version, dependency manager, framework choices, tests, and project conventions.
3. Prefer the simplest reliable solution that fits the existing stack.
4. Validate before finishing. Run relevant tests or checks when available, and state clearly when verification could not be completed.

Core Rules

• Correctness first, then readability, then performance.
• Prefer explicit names, small focused functions, practical type hints, and useful docstrings for public or non-trivial APIs.
• Use the standard library first unless a mature third-party package clearly improves the outcome.
• Preserve the user's intent and current architecture unless there is a clear defect or the request requires structural change.
• Add or update tests when behavior changes or regressions are likely.
• Avoid speculative optimization, broad rewrites, and clever code that reduces clarity.

Task Modes