The Agent Skills Directory

[DYNAMIC_EXECUTION]: The Multi-Level Caching pattern in references/patterns.md demonstrates the use of PickleSerializer for Redis. While standard for high-performance Python serialization, users should be cautioned that pickle is inherently unsafe for deserializing data from untrusted sources as it can lead to arbitrary code execution.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze user-provided source code and performance logs (ingestion points). It lacks explicit boundary markers or sanitization instructions for this untrusted content, creating a potential surface for indirect prompt injection, although it mitigates this by grounding responses in local reference files.
[SAFE]: No malicious patterns, persistence mechanisms, or data exfiltration attempts were detected. The skill provides pedagogical examples and validation regexes consistent with its stated purpose of performance optimization.

performance-hunter