The Agent Skills Directory

[SAFE]: The skill content is strictly pedagogical, offering advice on Pythonic coding standards, type hinting, and the use of the standard library. All examples provided are standard development patterns.
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process user-provided code for refactoring and debugging. While this involves processing untrusted data, the skill does not implement any dangerous capabilities, command execution, or network exfiltration paths that could be exploited via injection.
Ingestion points: User-provided Python source code and tracebacks for review (SKILL.md).
Boundary markers: None explicitly defined in the skill text.
Capability inventory: None; the skill provides guidelines for the agent's internal reasoning but does not include scripts or tools.
Sanitization: None mentioned.
[EXTERNAL_DOWNLOADS]: The skill mentions common Python development tools and libraries including ruff, black, mypy, pytest, and httpx. These are well-known, trusted industry tools. The code example using httpx is generic and does not reference any malicious or non-standard external domains.

python-expert