The Agent Skills Directory

[SAFE]: No malicious patterns, data exfiltration, or dangerous remote code execution detected. The skill's functionality is centered on codebase analysis and documentation.- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from a repository codebase. It mitigates risks by requiring the agent to ground answers in traceable code evidence and explicitly separate observed behavior from inference to avoid obedience to embedded instructions.- [COMMAND_EXECUTION]: Includes local Python scripts (scripts/validate.py, scripts/test_skill.py) for skill integrity and content validation. These scripts utilize standard library modules (os, sys, re, json, ast) for static analysis and do not perform network operations or dangerous system modifications.

product-question