Anthropic donates Agent Skills to the Agentic AI Foundation, under the Linux Foundation
The open standard that teaches AI agents specific tasks is moving to a neutral foundation. More than 70 platforms, including ChatGPT, Cursor and GitHub Copilot, have already integrated it.

An AI agent can write code or summarize documents with ease, but it falls short when asked to follow a company's exact procedure. Agent Skills were created to close that gap, and now Anthropic has announced that it is handing them over to a neutral foundation: the Agentic AI Foundation (AAIF), which operates under the umbrella of the Linux Foundation.
According to Anthropic, a skill is essentially a folder containing a SKILL.md file that teaches the agent a specific task or domain, along with any scripts or reference material it needs. The idea came about while the company was working on Claude Code, its programming tool. There, it noticed that the agent handled general work well but often lacked the procedural knowledge that real business tasks demand.
Adoption, according to the company, was fast. Within weeks of the October 2025 launch, developers had published thousands of skills on GitHub and written loaders (the components that allow skills to be read) for Codex, Cursor and OpenCode. In December 2025, Anthropic published Agent Skills as an open standard, with its own GitHub organization and its own domain, so that any agent could use it. Today, it says, more than 70 platforms have integrated it, including ChatGPT, Cursor, GitHub Copilot and VS Code; 1 million skills have been published and close to 280 million have been installed. These are the company's own figures, and the announcement does not explain how they are measured.
A neutral home for the standard
The donation includes the Agent Skills name and logo, the GitHub organization and the agentskills.io domain. The AAIF was co-founded last year by Anthropic, Block and OpenAI, and Anthropic had already donated the Model Context Protocol (MCP) to it. Those who want to use the standard or contribute to it do not need to join the foundation.
Maintenance will be in the hands of engineers from Anthropic, Google, OpenAI and other companies, who will participate in an individual capacity. David Soria Parra and Mahesh Murag will be the lead maintainers.
Anthropic proposes starting with a roadmap that includes numbered versions, a formal proposal process and security guidelines for those who write and install skills. That last point is no small matter: installing a skill means giving an agent instructions and potentially scripts, and until now there has been no common guidance on doing so. The project says that, as with MCP, it will prioritize community input and transparency in decision-making; feedback on changes to the specification is being collected on the project's Discord.
The announcement does not give dates for those versions or for the security guidelines. For developers in Latin America, the practical takeaway is that the specification is available at agentskills.io and the discussion takes place on GitHub, so a skill written once could work across multiple tools. It remains to be seen whether shared governance can maintain that compatibility as the standard evolves.
Source: Anthropic