Agents
Model Context Protocol (MCP)
Also known as MCP
MCP is an open standard for connecting LLM applications to tools and data sources.
It replaces bespoke, one-off integrations with a common interface any agent can reuse.
A single connector written once can serve different agents and applications instead of just one.
Adopting it early supports model-agnostic, vendor-neutral agent systems over tools wired to one assistant.
MCP standardizes the wire format, not the catalog, contract, or permission decisions behind it.
← Back to the full glossaryRelated terms
Tool Contract A tool contract is the schema a tool exposes to a model. MCP Gateway An MCP gateway is a shared layer in front of several MCP servers, centralizing controls. Tool Poisoning Tool poisoning hides an instruction in a tool's description that the model reads as legitimate. Confused Deputy A confused deputy is a component tricked into using more authority than a request deserves.
More in Agents
Agent (Agentic AI) An agent is an LLM given a loop, memory, and tools it can call. Tool Calling Tool calling lets a model request a structured action instead of only generating text. Context Engineering Context engineering decides what occupies a model's context window at every step, and in what order. Idempotency An idempotent operation produces the same result no matter how many times it runs.