"author:garagon" matching MCP servers:

Enables AI agents to locally scan untrusted third-party tool descriptions, configurations, and content for security threats before installing or trusting them, returning structured verdicts without network access.