The Agent Skills Directory

[SAFE]: The skill is a set of instructional prompts and local reference libraries for text processing. It does not perform network operations, file system modifications (beyond editing the provided text), or execute untrusted code.
[SAFE]: The disable-model-invocation: true flag indicates a restrictive configuration that limits the agent's ability to call other models or tools unless explicitly defined, reducing the attack surface.
[SAFE]: The reference files (references/patterns-zh.md and references/patterns-en.md) contain static text patterns and stylistic guidelines. No obfuscated scripts, Base64 strings, or hidden URLs were found within these files.
[SAFE]: The skill includes explicit instructions for the AI to protect the user's "Voice Sanctuary," specifically forbidding the fabrication of facts, data, or citations, which serves as a defense against hallucination-based data poisoning.

de-slop