The Register Four projects for facilitating AI agents move under vendor-neutral governance
The Agentic AI Foundation (AAIF) has been entrusted with stewardship of four additional AI projects in the hope of broadening the set of vendor-neutral tools for making AI agents.
The AAIF was founded in December 2025 under the auspices of the Linux Foundation as a vendor-neutral organization for the benefit of AI companies more interested in riding a tide that lifts all boats than making waves.
It launched with three projects under its supervision: Model Context Protocol (MCP), goose, and AGENTS.md. With the four new adoptees, there are now 10 projects sheltering under the AAIF's wings.
The donor organizations – Alibaba, Anthropic, Hugging Face, and OpenAI – have handed over control of OpenSandbox, Agent Skills, OpenEnv, and Open Responses respectively.
Each of these projects provides functions that developers of AI agents – which typically combine models, tools, and a task loop – may find helpful. And having a more or less agreed-upon blueprint for implementation tends to make deployment easier and more predictable, at least until your AI agent unexpectedly ransacks some third-party organization.
OpenSandbox provides a way to constrain or sandbox the capabilities of AI agents using various client SDKs, tools, protocols, lifecycle hooks, container runtimes, and assorted in-sandbox execution components.
Agent Skills outlines a format for text descriptions that guide AI agents through specified tasks more effectively. It provides context that helps agents accomplish their goals.
"We’re committed to keeping Agent Skills open, community-driven, and vendor-neutral," said Anthropic, an AAIF co-founder, in a blog post.
OpenEnv offers a containerized service for creating reinforcement learning environments that can be built without much trouble, can interoperate, and can scale to meet demanding testing.
And Open Responses is an open API specification for building LLM interfaces that can work with different model providers. It standardizes common request and response formats, streaming events, and tool invocation patterns.
The AAIF's nurturing of these technical specs has attracted considerable interest from companies convinced that there's a future in AI agents. The organization has surpassed 300 members in less than a year, with 60 coming on board this quarter, including Ant Group, Gates Foundation, MongoDB, Nutanix, Royal Bank of Canada, Vercel, and Vultr as Gold Members.
This enthusiastic (or credulous) embrace of AI agents may be a bit too rapid for some industries. Axios is reporting that US banks have become concerned that AI agents are interacting with customer accounts without bank authorization (though not necessarily without customer authorization). ®
Get AI news in your inbox
Daily digest of what matters in AI.
Key Terms Explained #
Agentic AI Agentic AI refers to AI systems that can autonomously plan, execute multi-step tasks, use tools, and make decisions with minimal human oversight.
AI Agent An autonomous AI system that can perceive its environment, make decisions, and take actions to achieve goals.
Anthropic An AI safety company founded in 2021 by former OpenAI researchers, including Dario and Daniela Amodei.
Hugging Face The leading platform for sharing and collaborating on AI models, datasets, and applications.