The Agentic AI Foundation (AAIF) has added four new projects to its roster, bringing the Linux Foundation-hosted organization to 10 hosted projects in its first year, up from the three it launched with in December 2025. The announcement came on October 22 during AGNTCon + MCPCon North America in San Jose, and marks one of the clearest signals yet that competing AI labs are willing to place shared agent infrastructure under neutral governance.

The four newcomers are Agent Skills, contributed by Anthropic; Open Responses, contributed by OpenAI; OpenEnv, contributed by Hugging Face; and OpenSandbox, contributed by Alibaba. Together, they cover four distinct layers of the agentic stack: how agents gain new skills, how applications talk to models across providers, how agents are trained and evaluated, and where agents run safely.

What Was Announced?

Each project addresses a different layer of the agentic AI stack. Agent Skills is an open format for giving agents new capabilities and domain expertise, packaging instructions, scripts and resources so they can be loaded on demand. Open Responses is an open specification for multi-provider, interoperable LLM interfaces, based on OpenAI’s Responses API, that defines a shared schema, consistent streaming events and extensible tooling. OpenEnv is an open framework and community hub for creating reinforcement learning environments, using a containerized service that exposes the agent interface, MCP and the Gym training protocol. OpenSandbox is a general-purpose sandbox platform for AI applications, with multi-language SDKs, unified sandbox APIs and Docker and Kubernetes runtimes for coding agents, GUI agents, evaluation and training.

Four layers of the agentic stack: Agent Skills, Open Responses, OpenEnv, OpenSandbox
Graphic: GPTOnline, using AAIF assets

Why It Matters

The new projects join the Agent2Agent (A2A) protocol, Agent Router, agentgateway, AGENTS.md, goose and the Model Context Protocol (MCP) under AAIF’s vendor-neutral governance. The contributor list is notable precisely because the companies behind it compete directly. Anthropic donated the Model Context Protocol to the foundation less than a year ago, and is now adding Agent Skills; OpenAI is contributing Open Responses after years of keeping its Responses API in-house.

“A year ago, AAIF launched with three projects and a bet that agentic AI needed a neutral home to thrive. Today we host 10,” said Manik Surtani, CTO and co-founder of the Agentic AI Foundation. “Agent Skills, Open Responses, OpenEnv, and OpenSandbox each cover a different part of that stack, from how agents gain capabilities to where they run, and four different companies chose to contribute them to neutral governance. That’s the job of this foundation: giving the community one place to build shared standards, protocols, and infrastructure, with no single company controlling the outcome in a way that puts the community first.”

David Soria Parra, co-creator and lead maintainer of MCP at Anthropic, framed the donation as a continuity move: “Without the right context, even the most intelligent agents fall short. We launched Agent Skills a year ago to give developers and enterprises alike an open standard to make their agents more capable. Today, we’re donating Agent Skills to the AAIF to ensure that anyone building with skills can shape what’s next.”

From OpenAI, Steve Coffey, member of technical staff, said Open Responses aims to help developers “build great products on today’s language models through simple, clean, and unified interfaces.” Hugging Face’s Ben Burtenshaw tied OpenEnv to the company’s open-stack philosophy: “Reinforcement Learning environments are a fundamental part of that stack because they teach agents the capabilities they will use in the real world.” Alibaba’s Xiaobin Xu, head of development infrastructure, said the company does not want OpenSandbox’s future “defined by a single company” and is creating “a path for more developers and organizations to help shape how sandbox environments work across the agent ecosystem.”

The announcement also points to where agentic AI infrastructure is heading. With MCP, AGENTS.md and now Agent Skills, Open Responses, OpenEnv and OpenSandbox under one roof, the foundation is positioning itself as the neutral home for the plumbing that lets agents work across platforms and providers. Whether cross-company adoption follows depends on governance and real implementations, not just the contributor list, but the direction is clear: the agent stack is being standardized in the open.

Sources