The Agent Skills Directory

[SAFE]: The skill is strictly diagnostic and read-only. It explicitly states in its metadata and instructions that it does not modify files and requires no network access, which eliminates the risk of data exfiltration or unauthorized system changes.
[SAFE]: No obfuscation, prompt injection, or remote code execution patterns were detected. The execution flow is transparently defined through a series of Markdown reference files that guide the agent through an audit process.
[SAFE]: The skill implements security best practices by encouraging the detection of vulnerabilities in the target codebase, such as hardcoded credentials or missing authentication, without introducing any such risks within its own implementation.

refactoring-audit