Once you really start using AI, you notice it quickly: you keep explaining how your company does things. What a quote looks like, which tone you use in emails, which steps go into a monthly report. That explanation is valuable. You want to write it down properly once and keep using it, even if you pick a different AI tool later.
That is exactly what the news of 22 October 2026 was about. Below is what was announced, with the sources. After that, our view on what you can do with it.
What was announced?
On 22 October 2026 Anthropic announced that it is donating Agent Skills to the Agentic AI Foundation (AAIF), which sits under the Linux Foundation. Anthropic is handing over the standard itself, plus the name, the logo, the GitHub organisation and the agentskills.io domain.
The AAIF confirmed it the same day: Agent Skills is one of four new projects at the foundation, alongside Open Responses (OpenAI), OpenEnv (Hugging Face) and OpenSandbox (Alibaba). The Linux Foundation published the same press release.
According to both Anthropic and the AAIF, more than 70 AI platforms already support the standard, over a million skills have been published and they have been installed nearly 280 million times. The specification is free to use under the Apache 2.0 and CC BY 4.0 licences.
What is a skill?
A skill is a folder containing a SKILL.md file. At the top of that file are a name and a short description of what the skill does and when to use it. Below that are the instructions. The folder can also hold scripts, reference documents and templates. (specification)
write-quote/
├── SKILL.md # name, description and the instructions
├── references/ # for example your price list or terms
└── assets/ # for example your quote templateWe made up this example for illustration.
An AI agent loads that knowledge in stages. At startup it reads only the name and description of every skill. Only when a task calls for it does it read the full instructions, and extra files only when they are needed. (agentskills.io) That way an agent can keep many skills on hand without becoming slow or confused.
Our view: what does this mean for small businesses?
This part is Forrict’s opinion, not part of the announcement.
Your way of working is the valuable part. You can replace the AI model. The knowledge of how your company works is not something you can simply buy. A skill is a place to capture that knowledge, in plain text, so your team and your AI do the same thing.
An open standard makes that knowledge less dependent. Now that the standard sits with an independent foundation, it is less likely that your instructions will only work in one product. That makes it easier to invest in them now.
Start small. Pick one task that comes back every week and always has to be done the same way. Capture it as a skill, test it with your team and improve it. Then move on to the next task.
How to get started
- List the tasks that come back every week.
- Pick one where everyone knows how it should go.
- Write down the steps the way you would explain them to a new colleague.
- Put that text in a skill and let your team work with it.
Want help with that first step? With AI with Claude we train your team and set up recurring workflows together with you.
Get in touch about AI in your company →
Sources
- Anthropic, 22 October 2026: Donating Agent Skills to the Agentic AI Foundation
- Agentic AI Foundation, 22 October 2026: Agentic AI Foundation Announces Four New Projects
- Agentic AI Foundation, 22 October 2026: Agent Skills: Open Standard for Sharing AI Agent Workflows
- Linux Foundation, 22 October 2026: press release on the four new projects
- Agent Skills: overview and specification
Forrict
AWS expert and consultant at Forrict, specializing in cloud architecture and AWS best practices for Dutch businesses.
