The Linux Foundation has assumed stewardship of the Model Context Protocol (MCP), marking a step toward standardizing how AI agents request data, execute tools, and access capabilities through well-defined APIs, reducing reliance on proprietary interfaces.
MCP, originally introduced as an open protocol for enabling structured, secure communication between AI systems and external resources, now gains a neutral home designed to support long-term governance and broad industry participation.
The move places MCP within the newly announced Agentic AI Foundation (AAIF), a consortium focused on building open standards and reference implementations for the rapidly expanding ecosystem of AI agents.
Alongside MCP, the foundation will also oversee related projects such as Block’s Goose and OpenAI’s AGENTS.md, ensuring that each component evolves in a coordinated, interoperable way.
According to the announcement, by bringing these efforts under one umbrella, the Linux Foundation aims to address fragmentation that has emerged as companies and researchers develop independent approaches to agent tooling.
With the Linux Foundation providing vendor-neutral oversight, now contributors from across the AI industry can collaborate on shared specifications without tied interests dictating direction. This governance model has previously supported critical technologies such as Kubernetes, SPDX, and the Open Container Initiative.
For more information, see GitHub’s announcement.
