The Agent Skills Directory

[SAFE]: The skill instructions and validation script are focused on document logic and structure, with no evidence of malicious patterns, obfuscation, or safety filter bypasses.
[REMOTE_CODE_EXECUTION]: The script does not use remote code or external dependencies. It relies on standard Python libraries and local modules for its validation logic.
[DATA_EXFILTRATION]: All data processing is local to the provided workspace. The skill reads draft outlines and writes summaries to an output directory, with no network access or data transmission detected.

argument-selfloop