· Luca Moretti
MCP explained: the plug between AI assistants and your own tools
Anthropic’s Model Context Protocol is now run by a Linux Foundation fund, has more than 10,000 servers in production and changed its core on 28 July 2026.
The Model Context Protocol (MCP) is an open standard that lets an AI assistant read from and act in other software: your CRM, your calendar, your database. Anthropic introduced it in November 2024; OpenAI adopted it in March 2025 and Google DeepMind followed.
Who runs it now
In December 2025 MCP was donated to the Agentic AI Foundation, a Linux Foundation fund co-founded by Anthropic, Block and OpenAI. By mid-2026 more than 10,000 MCP servers were in production and the SDKs were downloaded 97 million times a month.
The 28 July 2026 revision
The newest version makes the protocol stateless at its base layer. Two older features, sampling and roots, are deprecated and keep working for twelve months. If you run a connector, plan the update before mid-2027.
The risk to manage
Security researchers showed in April 2025 that prompt injection and poisoned tools can make an assistant leak data. Give a connector the narrowest access that does the job, log what it does, and keep a person in the loop for anything that sends money or e-mail.