Skip to content
Glossary

Model Context Protocol (MCP)

The Model Context Protocol (MCP) is an open standard, introduced by Anthropic, for connecting AI models to external tools and data sources through a uniform interface for tool discovery and invocation.

Definition

The Model Context Protocol (MCP) is an open standard, introduced by Anthropic, for connecting AI models to external tools and data sources through a uniform interface for tool discovery and invocation.

What MCP standardizes

Before MCP, every integration between a model and an external system was bespoke. MCP defines a common protocol so an agent can discover available tools and resources from an MCP server and invoke them through a consistent interface, the way a USB standard lets many devices speak to many hosts. It describes what an agent can do — the catalog of tools and how to call them — and is rapidly becoming the default connector layer for agent tooling.

MCP and governance

MCP standardizes connectivity but does not, by itself, govern it: an MCP server exposing a tool does not decide whether a given agent should be allowed to call it. That is where a control plane adds value. Cordum can gate MCP tool calls before dispatch — applying per-identity scope filters, requiring approval for specific tools, and recording each invocation — so the convenience of MCP comes with policy and audit. MCP describes what an agent can say; CAP and the control plane track what it actually did.

Frequently asked questions

Does MCP secure AI agents?

MCP standardizes how agents discover and call tools, but it does not decide whether a given call should be permitted. Securing MCP usage requires a governance layer that evaluates each tool call against policy before it executes and records it for audit.

How does MCP relate to CAP?

They are complementary. MCP describes what an agent can say — the tools available and how to invoke them. The Cordum Agent Protocol (CAP) carries what the agent actually did, so the platform can govern and audit behavior across the action lifecycle.

Related reading

Govern your AI agents with Cordum

Cordum is the agent control plane: policy-before-dispatch enforcement, human approvals, and a tamper-evident audit trail for autonomous AI agents.