mcp-guard is a security tool, so we take its own correctness seriously.

If you find a vulnerability in mcp-guard itself — for example a way to make the analyzer miss a poisoned description it should catch, or to break a consumer's build — please report it privately via GitHub Security Advisories rather than a public issue. We aim to acknowledge reports promptly.

mcp-guard performs build-time static analysis of MCP tool descriptions. It is one layer of defense and does not replace runtime protections — see the threat model for what is in and out of scope.