The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes several Python scripts (e.g., wikipedia-chatgpt-artifacts.py, wikipedia-communication-patterns.py) designed to be executed via a shell hook. These scripts perform static analysis on local project files using regular expressions to identify AI writing indicators. The execution is limited to reading project-specific directories (drafts, outlines, revisions, .planning) and does not involve network activity or access to sensitive system paths.
[SAFE]: The documentation and reference files contain links to established educational and research platforms such as Wikipedia, ArXiv, and the Association for Computational Linguistics (ACL). These references are relevant to the skill's purpose and do not point to suspicious or untrusted domains.
[SAFE]: Instructions in SKILL.md establish a logical workflow for content review and revision. While they use strong language ('The Iron Law') to enforce usage, the instructions remain within the bounds of content quality control and do not attempt to bypass agent safety filters or exfiltrate data.
[SAFE]: No obfuscation, data exposure, or persistence mechanisms were found across the analyzed files. The skill correctly identifies sensitive artifacts (like prompt refusals or internal model markers) as signs of AI generation for the user to review, rather than attempting to exploit them.

ai-anti-patterns