Run security checks against your own MCP server in CI, and see the findings in GitHub

Most MCP security tooling reads a server's tool descriptions. That finds text that looks suspicious....