The Agent Skills Directory

[SAFE]: A comprehensive audit of SKILL.md and all files in the references/ directory confirms the absence of prompt injection patterns, obfuscation techniques, or persistence mechanisms.
[SAFE]: The skill does not request or use any sensitive tools, nor does it perform file system writes, network requests, or command execution. Its operations are strictly limited to text analysis and restructuring.
[SAFE]: The instructional content regarding security (e.g., examples for password reset flows, rate limiting, and SQL injection sanitization in references/advanced_techniques.md) is provided solely as requirement templates for the user's own projects and does not involve any actual execution of those security functions or bypasses.
[SAFE]: The external link to WeChat (mp.weixin.qq.com) is a legitimate attribution for the methodology source and does not facilitate remote code execution or data exfiltration.

prompt-optimizer