How does Python Doctor calculate the health score?

The skill starts with a perfect score of 100 and deducts points based on the severity of confirmed findings—Critical (-10), High (-7), Medium (-5), and Low (-3)—across all audit categories.

How does the remediation feature work?

After an audit, you can trigger an optional fix loop. Claude will address the highest-impact findings first and then re-run the audit to report the improved health score.

Is it safe to run Python Doctor on untrusted code?

Yes. Python Doctor treats all repositories as untrusted by default and will only perform static analysis. Runtime checks like pytest or MyPy require explicit user approval before execution.

Does it support modern Python package managers?

Yes, it intelligently detects your environment and can use Poetry, uv, Pipenv, or standard Python runners depending on your project's lock files.

What specific Python issues can it detect?

It scans for 49 specific rules including SQL injection (Security), blocking async calls (Performance), mutable default arguments (Correctness), and circular imports (Architecture).

Python Doctor

Name: Python Doctor
Author: ragnarok22

byragnarok22

•

セキュリティとテスト

Audits Python codebases for security vulnerabilities, performance bottlenecks, and architectural antipatterns to provide a comprehensive health score and remediation plan.

Python Doctor is a sophisticated auditing tool for Claude Code designed to evaluate Python projects across four critical dimensions: Security, Performance, Correctness, and Architecture. By combining deep static rule scans with optional trusted runtime checks—including pytest, Ruff, and MyPy—it identifies high-impact issues ranging from SQL injection and blocking async calls to mutable default arguments and circular imports. The skill generates a deterministic 0-100 health score and a prioritized report, making it an essential utility for pre-release reviews, backend quality assessments, and automated codebase health monitoring.

主な機能

01Performance profiling for async blocking, inefficient loops, and memory management

02Deep security scanning for hardcoded credentials, injection risks, and unsafe deserialization

03Automated remediation loop to fix high-priority findings and track score improvements

04Deterministic 0-100 codebase health scoring across 49 specialized audit rules

053 GitHub stars

06Integration with trusted runtime tools like pytest, Ruff, and MyPy for validated checks

ユースケース

01Standardizing code quality checks during PR reviews or before deploying new microservices

02Identifying hidden performance bottlenecks and architectural technical debt in legacy projects

03Performing a comprehensive pre-release security and quality audit for Python backends

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ragnarok22/agent-skills python-doctor

For use in Claude.ai and ChatGPT

Download Skill