Agents · Fast-moving · Advanced

Model Context Protocol

Also known as: MCP

An open protocol for connecting AI assistants to external tools, data sources and prompts through a standard interface.

What Model Context Protocol is

MCP standardises the plumbing between models and systems. A server exposes resources, tools and prompts; any compatible client can use them without bespoke integration code.

How it works

Clients and servers exchange JSON-RPC messages over stdio or HTTP transports. Servers declare capabilities; the host application mediates which are exposed to the model and enforces permissions and consent.

Why it matters

It reduces the N-times-M integration problem in agent ecosystems, and has been adopted across multiple assistant and IDE vendors. Because servers can act on real systems, permission scoping and trust of third-party servers are the central security concerns.

Common uses

→Connecting assistants to databases and file systems
→IDE and developer tool integrations
→Reusable enterprise tool servers

Strengths

✓Write a connector once, use it in many clients
✓Clear capability boundaries

Watch for

✓Untrusted servers are a real risk
✓Specification and ecosystem still maturing

Continue exploring

Tool Use Function Calling AI API Prompt Injection