Anthropic has donated its Agent Skills standard to the Agentic AI Foundation, a standards body it co-founded with Block and OpenAI under the Linux Foundation umbrella. This move places the specification alongside the Model Context Protocol, which the company previously transferred to the same organisation. The action follows a prediction made by Simon Willison in December 2025, when he noted that open standards for agent capabilities would likely find a home in the AAIF.
The transfer signals a shift toward vendor neutrality in how artificial intelligence systems coordinate tasks. By moving the standard to an independent foundation, Anthropic removes itself as the sole gatekeeper for implementation details. This structure allows other developers to build compatible tools without relying on Anthropic’s specific infrastructure or licensing terms.
* The standard defines how AI agents request human assistance or external tools
* Block and OpenAI co-founded the Agentic AI Foundation last year
* The Model Context Protocol was previously donated to the same body
