MCP

MCP (Model Context Protocol) is an open protocol that standardizes how agents and LLM applications discover, connect to, and interact with external tool and data providers (called MCP servers).

Details

An MCP server exposes a set of tools (with schemas), resources (data the client can read), and prompts (reusable prompt templates) over a standardized transport (typically JSON-RPC over stdio or Streamable HTTP). From the model's perspective, MCP-served tools are function tools: the server declares each tool's name, description, and parameter schema, and the MCP client includes these schemas at inference time just like any other function tool. The agent runtime or host application acts as the MCP client, connecting to one or more servers and making their capabilities available to the model. This decouples function-tool implementation from the agent: any MCP-compatible server can be plugged into any MCP-compatible client without custom integration code.

Because MCP servers execute with the agent's permissions and their outputs feed directly into the model's context, they are a supply chain attack surface. A malicious or compromised MCP server can exfiltrate data, return poisoned results, or execute unauthorized actions.

Examples

• A file-system MCP server that lets a coding agent read and write project files through a standardized interface.
• A database MCP server that exposes query tools and schema resources to any connected agent.
• A browser-automation MCP server that provides navigation and interaction tools for web-based tasks.

Synonyms

Model Context Protocol