The Agent Skills Directory

[SAFE]: The skill operates entirely offline using Python scripts (detect.py and humanize.py) and a local JSON lexicon. No external network requests or data exfiltration attempts were detected.
[SAFE]: Static analysis flags for Unicode steganography are false positives. The scripts contain Unicode escapes and category checks (e.g., 'Cf') specifically designed to detect and strip invisible formatting characters while preserving legitimate emoji structures like zero-width joiners and regional flag tags.
[SAFE]: The tool includes logic to identify 'slop' vocabulary and machine-like structural patterns. These are implemented as standard regular expression replacements and reporting mechanisms with no dynamic execution or remote code loading.

x-human