The Agent Skills Directory

[SAFE]: The skill utilizes well-known, legitimate Python libraries such as aiocache, asyncpg, redis, aiohttp, and prometheus_client to implement its performance optimization features.
[PROMPT_INJECTION]: The skill includes an indirect prompt injection surface as it is designed to analyze user-provided Python source code for performance anti-patterns. 1. Ingestion points: User-provided code files processed via the validation rules in references/validations.md. 2. Boundary markers: None specified in the instructions for isolating code content. 3. Capability inventory: The skill's scope is restricted to performance analysis and code refactoring suggestions; it does not have access to sensitive system files or arbitrary network communication tools. 4. Sanitization: Not applicable as the skill performs static analysis using regular expressions rather than execution of the ingested code.
[REMOTE_CODE_EXECUTION]: The skill provides a pattern for using PickleSerializer with Redis caching in references/patterns.md. While the pickle module can be unsafe if used with untrusted data, its application here is for internal caching of application data, which is a standard practice for performance optimization in Python environments.

performance-hunter