12 Dec 2025 AI Engineering 5 min read

The Protocol Layer Goes Neutral

On December 9, 2025 the Linux Foundation announced the Agentic AI Foundation, co-founded by Anthropic, Block and OpenAI. Anthropic donated the Model Context Protocol, Block contributed goose, OpenAI added AGENTS.md. We examine what neutral governance changes for the agentic stack, what it deliberately leaves unsolved, and what engineering teams should do now.

A Standard Owned by One Vendor

Anthropic released the Model Context Protocol on November 25, 2024, as an open standard for connecting AI models to tools and data. MCP defines a JSON-RPC-based client-server protocol: a model-side client discovers and calls tools, resources and prompts that servers expose. Twelve months later the numbers are unambiguous: more than 10,000 published MCP servers, official SDKs with over 7 million monthly downloads across Python and TypeScript, and support in Claude, ChatGPT, Gemini, Microsoft Copilot, Cursor and VS Code.

De facto is not the same as neutral. Until this week the specification, the repositories and the trademark belonged to Anthropic. Every contribution from Google, Microsoft or Block improved a competitor's property, and every roadmap decision ultimately sat with one company. Nothing but goodwill prevented a future change of license or direction. For a library that risk is acceptable. For infrastructure an industry standardizes on, it is a structural defect.

An agent reaches its tools over MCP — one standard surface. 1/4

What Happened on December 9

On December 9, 2025, the Linux Foundation announced the formation of the Agentic AI Foundation (AAIF), co-founded by Anthropic, Block and OpenAI with support from Google, Microsoft, AWS, Cloudflare and Bloomberg. Eight Platinum members anchor the fund; Gold and Silver members range from Cisco and IBM to Hugging Face and Zapier. Three projects were donated at launch: MCP, goose and AGENTS.md.

The operative word is donation, not partnership. Ownership transfers to the foundation. A directed fund is the Linux Foundation's structure for an industry consortium with its own budget, membership and technical governance under the Foundation's legal umbrella — the same organization that stewards the Linux kernel, Kubernetes, Node.js and PyTorch. Anthropic states that MCP's governance model remains unchanged: the same maintainers, the same community-driven decision process.

Three Donations Define the Stack

The three founding projects are not arbitrary. They cover three distinct layers of an agentic system: MCP standardizes how agents reach tools and data, goose is a working open-source agent that consumes that protocol, and AGENTS.md standardizes how repositories instruct coding agents. Together they form a minimal, fully open reference stack.

None of the three is a model. That is the point. The AAIF governs the connective tissue — protocol, runtime, convention — while the models above it remain proprietary and competitive. The foundation draws a line through the stack: below it, cooperation; above it, competition.

Project	Donor	First released	Layer
MCP	Anthropic	November 2024	Protocol: connects models to tools and data; 10,000+ published servers
goose	Block	Early 2025	Runtime: local-first open-source agent framework built on MCP
AGENTS.md	OpenAI	August 2025	Convention: per-repository agent instructions; 60,000+ adopting projects

Why Neutrality Is an Engineering Concern

Neutral governance changes three concrete things. Intellectual property: specification, repositories and trademarks are held by the foundation, and no single member can relicense them. Process: changes follow documented governance instead of one vendor's roadmap. Incentives: OpenAI engineers can now improve MCP without strengthening an Anthropic asset, and vice versa.

The pattern has precedent. Google moved Kubernetes into the newly created CNCF in 2015; the project became an industry default partly because no competitor had to build on Google's property. The Node.js Foundation, also 2015, ended the io.js fork. Foundation transfer is the normal end state of successful open infrastructure. MCP reached it in just over a year.

For enterprise buyers, neutrality is a procurement criterion, not a sentiment. A standard held by a foundation survives the strategy changes of any single company. That is the difference between adopting a protocol and adopting a product.

What the Foundation Does Not Solve

The donation changes ownership, not technology. MCP's open problems remain open. Prompt injection through tool descriptions and server responses is unsolved. The quality of the 10,000+ published servers varies widely, and no certification separates sound implementations from abandoned ones. Authorization improved across the 2025 spec revisions, but securing deployments remains the operator's job.

AGENTS.md standardizes where instructions live, not how agents interpret them; two agents reading the same file can still behave differently. Governance also has a price: foundation consensus is slower than single-vendor decisions, and the Linux Foundation hosts stalled projects alongside thriving ones. Neutrality buys stability. It does not buy velocity.

Fragmentation above the protocol also persists. Orchestration frameworks remain incompatible; the AAIF standardizes how agents reach tools, not how teams build, evaluate or supervise agents. Anyone expecting a full-stack standard will be disappointed. The foundation governs plumbing, deliberately.

Consequences for Engineering Teams

The risk calculus for adopters is now simpler. An integration written once as an MCP server serves every compliant client, and the protocol underneath it can no longer be unilaterally relicensed. The practical consequences: expose internal systems as MCP servers instead of per-platform plugins, maintain an AGENTS.md in every repository, and keep orchestration logic cleanly separated from protocol bindings — the layers above the protocol remain proprietary and volatile.

We at Blue IT Systems standardized client integrations on MCP in the course of 2025. The residual risk in that decision was governance, not technology: a protocol owned by one vendor can change underneath you. Since December 9 that risk is materially smaller, and we see no remaining reason to hedge with parallel integration paths.

Outlook

The near-term roadmap is visible. The specification release of November 25, 2025 — asynchronous operations, statelessness, server identity, official extensions — will propagate through the SDKs in early 2026. From the AAIF we expect what young standards need next: a conformance test suite, security profiles for enterprise deployment, maturation of the official server registry, and further project donations.

Two predictions. First, by the end of 2026 MCP support will be a checkbox, not a differentiator; competition moves up the stack to orchestration, evaluation and safety layers. Second, "MCP-compatible" becomes a testable claim once a conformance program exists. The open question is pace: foundations guarantee that a standard survives, not that it stays ahead. The next twelve months will show whether neutral governance can move at the speed agentic AI demands.