Somewhere, a very humble text file called SKILL.md is having the best week of its life. It started as an internal trick inside a coding tool, got adopted by basically everyone, and has now been handed to a neutral foundation like a prized family recipe going to the town cookbook.
From Side Project to Shared Standard
On October 22, Anthropic announced it is donating Agent Skills, the open standard for packaging know-how for AI agents, to the Agentic AI Foundation (AAIF), which operates under the Linux Foundation. A skill is just a folder containing a SKILL.md file plus any scripts or reference material an agent needs to do a specific task. Anthropic is handing over the name, logo, GitHub organization and the agentskills.io domain.
The adoption numbers explain why this matters. More than 70 agent platforms have integrated Agent Skills, including ChatGPT, Cursor, GitHub Copilot and VS Code, with about 1 million skills published and nearly 280 million installs. Engineers from Anthropic, Google, OpenAI and other companies will maintain the spec as individuals, and AAIF, co-founded last year by Anthropic, Block and OpenAI, already houses the Model Context Protocol.
Boring Governance, Exciting Consequences
Standards bodies rarely trend, but this is the plumbing that decides whether your AI investments are portable. When one vendor controls a format, every skill you write is a small bet on that vendor. When a neutral foundation controls it, a skill your team writes for one agent has a fighting chance of working in the next one, which is precisely what businesses wary of lock-in have been asking for.
The bit most people will skim past is on the roadmap: security guidance for skill authors and installers, alongside versioned releases and a formal proposal process. With hundreds of millions of installs, skills are effectively a new package ecosystem, and if npm taught us anything, it's that popular package ecosystems attract people with bad intentions. A skill can bundle scripts, so "install this handy skill" deserves the same suspicion as "run this random shell script."
The agent wars aren't over, but everyone just agreed on the shape of the ammunition, and that's a bigger deal than it sounds.
Thinking about capturing your company's processes as reusable skills for AI agents? We help businesses design AI integrations that stay portable as the standards keep shifting.
Source: Anthropic