Agentic AI Foundation: MCP, AGENTS.md and goose move to the Linux Foundation
2025-12-09 · Publisher : Linux Foundation · Domain : Artificial Intelligence
The Linux Foundation creates the Agentic AI Foundation (AAIF), home to Anthropic's Model Context Protocol, OpenAI's AGENTS.md and Block's goose: the open building blocks of agents move under neutral governance.
What changes
- Three founding projects: MCP (Anthropic), AGENTS.md (OpenAI) and goose (Block).
- More than 10,000 published MCP servers and more than 60,000 open source projects using AGENTS.md, according to the Linux Foundation.
- Platinum members: AWS, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft and OpenAI.
- Direct competitors jointly fund neutral governance for agent standards.
- For companies, MCP becomes a durable standard rather than a single vendor's technology.
On December 9, 2025, the Linux Foundation announces the creation of the Agentic AI Foundation (AAIF), a foundation meant to host the open projects that structure the AI agent ecosystem. Three founding contributions, made by competing companies, launch it. ### What happened - **Model Context Protocol (MCP)**: released by Anthropic in November 2024, the protocol standardizes how models connect to tools, data and applications. The Linux Foundation counts more than 10,000 MCP servers, from developer tools to enterprise deployments. - **AGENTS.md**: proposed by OpenAI, this Markdown file gives coding agents the instructions specific to a repository (commands, conventions, tests). It is used by more than 60,000 open source projects and supported by tools such as Cursor, GitHub Copilot and VS Code. - **goose**: Block's open source agent, local and extensible, built on MCP integrations. - **Members**: eight platinum members (AWS, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft, OpenAI), eighteen gold members including Cisco, Datadog, Docker, IBM, JetBrains, Okta and Oracle, and more than twenty silver members. ### Why it matters In one year, MCP went from an Anthropic proposal to a standard adopted by its competitors. Handing it to a neutral foundation, alongside AGENTS.md, ensures no single vendor can change its rules alone. Much as Kubernetes did for the cloud, agents now have a common open foundation, with shared governance and roadmap. ### Caveats Open governance does not solve agent-specific security issues (prompt injection, MCP server permissions, supply chain), which remain the responsibility of those who deploy them. ### Key takeaways At the end of 2025, agent standards (tool connection, project instructions) move beyond a single vendor: a sign of maturity and longevity for companies investing in these building blocks.
Links
Tags : IA, Agents, Open Source, MCP, Standards ouverts, Gouvernance