The Agent Skills Directory

[SAFE]: The skill serves a legitimate utility purpose for development workflows and contains no code obfuscation, remote execution, or privilege escalation patterns.
[COMMAND_EXECUTION]: The instructions recommend using standard search utilities like grep to verify the codebase for existing implementations, which is a benign and intended part of the skill's logic.
[DATA_EXFILTRATION]: The TypeScript implementation checks for the existence of common project files (e.g., README.md, CLAUDE.md) using standard Node.js APIs to verify project context. These operations are read-only and limited to metadata verification with no network activity detected.

Confidence Check