What is the Model Context Protocol (MCP)?

The Model Context Protocol (MCP) is an open specification for how AI agents and development tools talk to external tools and data. An agent connects to an MCP server. The server exposes tools the agent can call and resources it can read.

Key Concepts

• MCP Server: An application that implements the MCP specification. It exposes a set of tools and resources that can be used by other MCP-enabled applications.
• Tools: Functions that can be executed by an MCP server. For example, a tool might create an issue in a bug tracker or search for a document in a knowledge base.
• Resources: Data sources that can be accessed through an MCP server. This could be anything from a file on the local system to the response from a third-party API.

Why is MCP Important?

MCP enables a new level of interoperability between AI agents and developer tools. By providing a standardized communication protocol, it allows for the creation of powerful, integrated workflows that were not previously possible.

For example, an AI coding assistant could use an MCP server to:

• Access documentation from a knowledge base
• Create and update issues in a project management tool
• Search for relevant code snippets in a shared repository

This allows the AI to have a much deeper understanding of the development environment and to perform more complex tasks with greater accuracy.

For teams running MCP-connected agents in production, interoperability also creates a governance problem: any tool the agent can reach, it can call. Preloop sits in the MCP path so teams can set allow, deny, and approval rules, attribute model spend, and keep an audit trail. See EU AI Act readiness with Preloop.