Anthropic, the company behind the Claude chatbot, said on October 22 that it is giving away control of Agent Skills, a format it created for teaching AI agents how to do specific jobs. The standard now belongs to the Agentic AI Foundation, a group that runs under the non-profit Linux Foundation. Anthropic says the move gives the standard “a neutral home” so no single company decides its future.
What are Agent Skills?
An AI agent is a chatbot that can take actions for you, such as editing files or working through a multi-step task. General knowledge only gets an agent so far. To do real work, it also needs to know how a particular company or task does things.
A “skill” fills that gap. In simple terms, it is a folder containing an instruction file (called SKILL.md) that tells the agent how to handle one kind of task. It can also hold scripts or reference documents. Anthropic says it came up with the idea while building Claude Code, its coding assistant, because the agent was good at general work but lacked the step-by-step know-how that business tasks need.
How widely is it used?
Anthropic launched skills in October 2025 and published them as an open standard in December 2025. According to the company:
- More than 70 agent platforms now support the format, including ChatGPT, Cursor, GitHub Copilot and VS Code.
- About 1 million skills have been published.
- Those skills have been installed nearly 280 million times.
That spread across rival products is a big reason for the handover. When competitors such as OpenAI and Google rely on a format, they are more comfortable if it is not owned by one of their rivals.
What exactly changes hands
Anthropic is transferring the Agent Skills name and logo, its GitHub organisation (where the code and specification live) and the agentskills.io website. Engineers from Anthropic, Google, OpenAI and other companies will look after the standard as individuals rather than as company representatives. Two people, David Soria Parra and Mahesh Murag, have been named lead maintainers.
The group has proposed next steps that include numbered releases, a formal process for suggesting changes, and security guidance for people who write and install skills. That last point matters, because a skill can include scripts that run on your computer.
This is not the first time Anthropic has done this. The Agentic AI Foundation was co-founded in 2025 by Anthropic, Block and OpenAI, and Anthropic previously donated the Model Context Protocol (MCP), a widely used standard for connecting AI tools to other apps and data.
What it means for you
For most people, nothing changes day to day. Anthropic says no action is needed to keep using skills, and anyone can keep using or contributing to the standard without joining the foundation. The format itself is free and open.
The longer-term benefit is that a skill written once should keep working across different AI assistants, not just Claude. If you use agent tools at work, the promised security guidance is worth watching: only install skills from sources you trust, since they can contain code.