Agent tool protocol
Model Context Protocol: where it fits.
Model Context Protocol, or MCP, is an open protocol for connecting AI applications to tools, data sources, and external systems.
Best fit: Use MCP when multiple agent tools need a common way to expose files, APIs, databases, or services to an AI assistant.
Strengths
- Useful connective tissue for agent ecosystems.
- Helps avoid one-off custom integrations for every assistant.
- Increasingly common across coding agents and AI apps.
Watch-outs
- A protocol is not a security model by itself.
- Only connect servers you trust and understand.
- Treat tool descriptions, prompts, and external data as part of the attack surface.