Anthropic donated Agent Skills, the open standard that lets an AI agent load a folder of instructions for a specific task, to the Agentic AI Foundation on October 22, 2026. The foundation sits under the Linux Foundation; Anthropic co-founded it with Block and OpenAI, and it already holds the Model Context Protocol.
The numbers in Anthropic’s announcement: more than 70 agent platforms have integrated the standard since it was published in December 2025, 1 million skills have been published and nearly 280 million installed.
On this page · 7 sectionsOpen
Anthropic donated Agent Skills, the open standard for teaching AI agents a task through a folder with a SKILL.md file, to the Agentic AI Foundation on October 22, 2026.
The name and logo, the GitHub organization and the agentskills.io domain move to the foundation, which sits under the Linux Foundation and already holds MCP.
Anthropic reports more than 70 agent platforms using the standard, including ChatGPT, Cursor, GitHub Copilot and VS Code, with 1 million skills published and nearly 280 million installed.
Engineers from Anthropic, Google, OpenAI and other companies maintain it as individuals, led by David Soria Parra and Mahesh Murag.
The proposed roadmap starts with versioned releases, a formal proposal process and security guidance for people who write and install skills.
§ 01What changes hands
Anthropic is giving the foundation everything that made the standard its own, and keeping a seat as one maintainer among several.
| Item | Held by | From October 22, 2026 |
|---|---|---|
| Agent Skills name and logo | Anthropic | Agentic AI Foundation |
| GitHub organization | Anthropic | Agentic AI Foundation |
| agentskills.io domain | Anthropic | Agentic AI Foundation |
| Maintainers | Anthropic | Engineers from Anthropic, Google, OpenAI and others, as individuals |
| Lead maintainers | Not named | David Soria Parra and Mahesh Murag |
Nothing changes for people who use skills today. Anthropic’s words: “You can continue to use the standard or contribute to it without joining the AAIF.” The specification stays where it was, and the project’s GitHub organization moves with it.
§ 02What a skill is
Anthropic’s definition: “A skill is a folder containing a SKILL.md file that teaches an agent a specific task or domain, along with any scripts or reference material it needs.” The agent reads the short description up front and opens the rest only when the task comes up, so one agent can carry hundreds of skills without loading them all.
Skills came out of Claude Code. Anthropic says the agent “handled general knowledge work well, but often lacked the procedural knowledge that real business tasks need.” Developers then wrote loaders for Codex, Cursor and OpenCode within weeks, which is what pushed Anthropic to publish the format as an open standard.
§ 03From a Claude Code feature to 70 platforms
| Measure | Figure |
|---|---|
| Agent platforms integrated | More than 70 |
| Platforms Anthropic names | ChatGPT, Cursor, GitHub Copilot, VS Code |
| Skills published | 1 million |
| Skills installed | Nearly 280 million |
These are Anthropic’s figures; no independent count exists yet. The installed number counts installs, not distinct users.
§ 04What comes next
Anthropic has proposed a starting roadmap: versioned releases, a formal proposal process, and security guidance for people who write and install skills. Feedback on specification changes goes through the project’s Discord.
David Soria Parra, one of the two lead maintainers, wrote on X: “I am so proud that we are taking the most impactful standard for Agents in the last years and donate it to the Agentic AI Foundation.”
The security item is the one to watch. A skill can include scripts, and with nearly 280 million installs, guidance on vetting what you install is still a roadmap item rather than a published rule.
§ 05What it means if you use skills
- A skill you write now is less tied to one vendor; the format’s direction is set in a shared project.
- Watch for the first versioned release, which tells you what to pin when tools disagree.
- Treat a third-party skill like code you install: read its scripts before an agent runs them.
We build AI employees, and CellCog’s employees keep their methods as skills in this same folder and SKILL.md shape, so a neutral home for the format matters to us directly. We have no role in the foundation.
§ 06Update log
- October 22, 2026: page opened from Anthropic’s announcement and the lead maintainer’s post.
- October 22, 2026: explainer scene added.
§ 07Sources
- Anthropic, Donating Agent Skills to the Agentic AI Foundation, October 22, 2026.
- agentskills.io, the specification.
- David Soria Parra, post on X, October 22, 2026.
Q1What is an Agent Skill?
A folder containing a SKILL.md file that teaches an agent a specific task or domain, plus any scripts or reference material it needs. Agents load it when the task comes up.
Q2What is the Agentic AI Foundation?
A foundation under the Linux Foundation that Anthropic co-founded with Block and OpenAI. It already holds the Model Context Protocol, which Anthropic donated earlier.
Q3Who maintains Agent Skills now?
Engineers from Anthropic, Google, OpenAI and other companies, acting as individuals, with David Soria Parra and Mahesh Murag as lead maintainers.
Q4Do existing skills still work?
Yes. The donation moves ownership and governance; Anthropic announced no change to the format itself.
Q5Are third-party skills safe to install?
Treat them like code you install: a skill can carry scripts. Security guidance for writing and installing skills is on the proposed roadmap, not published yet.