Model Context Protocol
Protocol for Tool learning
- Hosts are LLM applications (like Claude Desktop or IDEs) that initiate connections
- Clients maintain 1:1 connections with servers, inside the host application
- Servers provide context, tools, and prompts to clients

MCP Usages
Limitation
The structure requiring code execution from users poses security risks and is vulnerable to malicious inputs. There are authentication issues and no restrictions on dangerous tools. It's difficult to implement interactive UX and is vulnerable to AI Hacking techniques like Prompt Injection.
Docs with modelcontextprotocolmodelcontextprotocol • Updated 2025 Jul 19 23:6
modelcontextprotocol
modelcontextprotocol • Updated 2025 Jul 19 23:6
from Anthropic
review
roadmap
What it means compared to, Network API