The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local utility scripts (profile_store.py, validate_skill.py) for managing configuration and validation. These scripts use standard Python modules like argparse, json, and pathlib, and follow best practices such as atomic writes and restricted file permissions.\n- [CREDENTIALS_UNSAFE]: The profile_store.py script includes explicit logic to filter out sensitive keys (e.g., tokens, secrets, passwords) before persisting data to the user profile, reducing the risk of accidental credential leakage during profile updates.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided content for file storage. Instructions in SKILL.md require the agent to use specific serializers and validators for structured data (JSON, YAML, CSV) and to preserve the original text verbatim, which serves as a mitigation against data-driven instruction override.\n- [REMOTE_CODE_EXECUTION]: The skill specifies a dependency on the PyYAML library for validation purposes. This is a standard library dependency used within a controlled validation script and does not involve downloading or executing code from unverified remote sources.

lov-output