MCP (Model Context Protocol)
MCP (Model Context Protocol) is a standard protocol that enables AI agents to connect to external tools, databases, and APIs. It is an open standard developed by Anthropic and donated to the Linux Foundation's Agentic AI Foundation.
Even as AI agents grow more sophisticated, they remain little more than "eloquent conversation partners" if they cannot connect to the outside world. MCP is the industry's answer to this challenge, establishing a common communication protocol between agents and external tools.
For example, by connecting a Supabase MCP server, an agent can invoke SQL execution, migration application, and type generation as tools. With a GitHub MCP server, it becomes possible to read Issues and create PRs. The agent side dynamically discovers "which tools are available" via the protocol and calls them as needed.
In terms of positioning, MCP handles "vertical" integration — the connection from agents to the external world. Meanwhile, "horizontal" communication between agents is handled by A2A (Agent-to-Agent Protocol), meaning MCP and A2A are complementary rather than competing.
In December 2025, Anthropic, together with OpenAI and Block, established the Agentic AI Foundation (AAIF) within the Linux Foundation and donated MCP to it. As an open standard designed to avoid vendor lock-in, it has been adopted by numerous tools including Cursor, Windsurf, and Claude Code.
Related Terms
A2A (Agent-to-Agent Protocol)
A2A (Agent-to-Agent Protocol) is a communication protocol that enables different AI agents to perform capability discovery, task delegation, and state synchronization, published by Google in April 2025.
Base Model (Foundation Model)
A base model (Foundation Model) is a general-purpose AI model pre-trained on large-scale datasets. Rather than being specialized for a specific task, it functions as a "foundation" that can be adapted to a wide range of applications through fine-tuning or prompt engineering.
Context Engineering
Context Engineering is a technical discipline focused on systematically designing and optimizing the context provided to AI models — including codebase structure, commit history, design intent, and domain knowledge.
