The Agentic AI Foundation added four projects, bringing it to 10 in its first year

The Agentic AI Foundation, the Linux Foundation backed home for open agent standards, announced four new hosted projects on October 22: Agent Skills from Anthropic, Open Responses from OpenAI, OpenEnv from Hugging Face and OpenSandbox from Alibaba. That brings it to 10 projects in its first year, up from three when it launched in December 2025.

The announcement came at AGNTCon and MCPCon North America in San Jose. The new projects join the Model Context Protocol, the Agent2Agent protocol, AGENTS.md, goose, agentgateway and Agent Router under the foundation's vendor neutral governance.

Each new project covers a different layer of the agent stack. Agent Skills is about how agents pick up new capabilities, packaged as reusable instructions and resources. Open Responses is meant to give developers one common API format for talking to models across providers. OpenEnv covers how agents are trained and evaluated in shared environments. OpenSandbox addresses where agents can run code safely.

"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, the foundation's CTO and co-founder.

Why it matters for developers: the big labs are handing their agent plumbing to a neutral body, which lowers the risk of building on it. Open Responses is the one to watch most closely. A shared request and response format across providers would make it easier to swap models, route between them and avoid lock-in. Agent Skills under neutral governance means skills written for one agent are more likely to work in others.

What's next: look for the specs and reference implementations to land in public repositories, for other model providers to adopt Open Responses, and for tooling that supports skills across different agents.

Building with AI? See this week's hot AI repos and how to use each one.