Anthropic hands Agent Skills to a neutral foundation, with 280 million installs

Anthropic is donating Agent Skills, its open standard for giving AI agents specialized knowledge and workflows, to the Agentic AI Foundation (AAIF), according to its announcement of October 22. More than 70 agent platforms have integrated it.

The move, also picked up by Techmeme, follows Anthropic’s earlier donation of the Model Context Protocol (MCP) to the same foundation.

What moves to the foundation, and who runs it

The AAIF sits under the Linux Foundation. Anthropic says the donation covers the standard itself, the Agent Skills name and logo, the GitHub organization and the agentskills.io domain. In practice, that means the project’s identity no longer belongs to one company.

A skill, as Anthropic describes it, is a folder containing a SKILL.md file plus any scripts or reference material the task needs. Skills were first built for Claude Code and launched in October 2025. Anthropic published the standard in December 2025.

On governance, engineers from Anthropic, Google, OpenAI and other companies will maintain the standard as individuals rather than as company representatives. David Soria Parra and Mahesh Murag are the Lead Maintainers. Anthropic says the project will follow MCP’s approach, with community input and transparent decision-making, and that anyone can use or contribute to the standard without joining the AAIF.

Adoption figures behind the handover

Anthropic lists ChatGPT, Cursor, GitHub Copilot and VS Code among the platforms that have integrated the standard since December. It also reports 1 million skills published and nearly 280 million installs. These are the company’s own figures, and the announcement does not say how installs are counted or verified.

The proposed first priorities are versioned releases, a formal proposal process, and security guidance for people who write and install skills. These are plans, not shipped features, and no dates are given. The announcement states no license terms, so readers who build on the standard should check the repository at github.com/agentskills/agentskills.

A neutral home for a standard rivals already use

The angle here is rules and safety, because the open question is trust in shared skills. Skills can carry scripts, and a standard that makes sharing easy also makes it easier to inherit a skill from a stranger. One analysis cited in coverage of the donation raises this point. Anthropic’s own roadmap names security guidance as a priority, but it has not published that guidance yet.

The foundation is growing quickly. According to a trade report on the AAIF’s announcement, Agent Skills is one of four new projects, alongside Open Responses, OpenEnv and OpenSandbox, bringing the foundation to 10 hosted projects in its first year. It launched in December 2025 with three. Anthropic co-founded it with Block and OpenAI, so the donated standard now sits under a body that two of the biggest model makers helped create.

The timing also fits a wider push to give agents controlled access to real accounts. Our recent coverage of 1Password’s plan to let AI agents sign in without seeing the password and of ServiceNow’s Arc agent shows vendors competing on how safely agents act. A shared skills format addresses a different layer, what agents know how to do, but the same trust questions apply to both.

What changes for developers now

Anthropic says existing skills keep working. Nothing in the announcement requires developers to change code today. What shifts is who decides the standard’s direction: proposals will go through a process the foundation’s maintainers run, instead of one vendor’s roadmap.

For teams that write skills once and run them across several tools, a vendor-neutral home reduces the risk that one company changes the format. That is the benefit Anthropic claims, and it will only be proven if the promised versioning and proposal process arrive and other companies’ maintainers shape the result. Feedback on spec changes goes through the project’s Discord and the GitHub discussions.

Anthropic did not say whether it will keep funding the project, or what it expects the maintainers to change first.