The Model Context Protocol is an open specification that lets AI applications, including AI agents, discover and call external tools and data sources in a standardized way. Because MCP servers frequently expose sensitive actions and data, the protocol's authorization model builds directly on OAuth 2.0, which is why MCP and agentic identity are so closely linked.
Frequently asked questions
Is MCP an identity standard?
Not by itself, it's a tool-calling protocol, but its authorization model relies on OAuth 2.0, which is why it appears throughout agentic identity discussions.
Does MCP replace OAuth 2.0?
No, MCP builds on top of it rather than replacing it; OAuth 2.0 still governs how an MCP client proves it has permission to call a given tool.
Was this definition useful?
Feedback goes to the editor who maintains this entry.
.avif)