Dredd MCP

A preflight security layer that checks MCP servers before your agent invokes their tools. Dredd queries its threat intelligence backend to flag compromised dependencies, tool surface drift, remote URL changes, and permission escalation. It returns BLOCK, ADVISORY, or ALLOW verdicts in under 200ms, each HMAC-signed. You call check_mcp_server with a server name, optional version, and optional tool name before execution. Useful if you're consuming third-party MCP servers and want a second opinion on whether their dependency tree or runtime behavior has shifted since you last vetted them. Fails open by design so it won't brick your workflow if the backend is unreachable.