MCP Registry
What is MCP Registry?
The MCP Registry is the official, centralized metadata directory at registry.modelcontextprotocol.io for publicly accessible MCP servers, verifying who published each one through namespace authentication rather than scanning what the server's code actually does. AdminKit's guide to securing an MCP server covers why a server's listing there is not, by itself, a safety signal worth trusting.
By AdminKit team, Netbase JSC · Updated
Anthropic introduced the Model Context Protocol as an open standard for connecting AI systems to external tools and data, and the MCP Registry grew out of that same effort: it is the official, community-run metadata directory for publicly accessible MCP servers, publishing a standardized server.json record (name, install location, declared capabilities) for each one so MCP clients and downstream marketplaces can discover them. Its own documentation is explicit about what that listing does and doesn’t mean: namespace authentication ties a server’s reverse-DNS-style name to a verified GitHub account or domain, proving who published it — the registry’s own implementation is itself open-sourced on GitHub — while actual security scanning is delegated entirely to the underlying package registry (npm, PyPI, Docker Hub) or to downstream aggregators that choose to add it. The registry’s own moderation policy says so directly: it removes malware, spam and illegal content, but explicitly does not remove a server just for having security vulnerabilities or being low quality. That gap is exactly what makes tool poisoning effective against a server pulled from the registry without further review — authenticity and safety are two different guarantees, and the registry only makes the first one.
MCP Registry is part of AdminKit’s AI agents & MCP cluster; the full AdminKit glossary lists other related terms.
Frequently asked questions
Does a listing in the MCP Registry mean a server is safe to use?
No -- namespace authentication only proves who published a server, tying its name to a verified GitHub account or domain; actual security scanning is delegated to the underlying package registry or downstream aggregators.
Does the MCP Registry remove servers with security vulnerabilities?
No -- its own moderation policy says it removes malware, spam and illegal content, but explicitly does not remove a server just for having vulnerabilities or being low quality.
What does the MCP Registry actually verify?
Namespace authentication: that a server's reverse-DNS-style name is tied to a verified GitHub account or domain, proving who published it -- not what the server's code actually does.
Why doesn't a registry listing prevent tool poisoning?
Because authenticity and safety are different guarantees; a registry-listed server can still carry malicious or poisoned tool descriptions since the registry doesn't scan for that.