Run this before you deploy any MCP server and it'll scan for compliance gaps against EU AI Act Article 50 requirements, plus general security and quality issues. It checks your tool definitions for completeness, validates OAuth flows and token handling, and flags common security misconfigurations. Think of it as a preflight checklist that catches the regulatory and security footguns you'd otherwise miss until production. Especially useful if you're shipping tools that handle European user data or need to document AI system transparency requirements. The scanner runs via streamable HTTP, so you can integrate it into CI pipelines or hit it manually before releases.
claude mcp add --transport http io.github.shelvick-mcp-compliance-preflight https://stream-4-mcp-compliance-preflight.dynsup.workers.dev/mcp