Anthropic donates Agent Skills to the AAIF

On 22 October 2026 Anthropic announced it is donating Agent Skills—the open standard that gives AI agents specialised knowledge and workflows—to the Agentic AI Foundation (AAIF) under the Linux Foundation. More than 70 agent platforms have integrated the standard since December 2025, with 1 million skills published and nearly 280 million installs. Full details in Anthropic’s announcement: https://www.anthropic.com/news/donating-agent-skills-to-the-agentic-ai-foundation.

Key takeaways

• On 22 October 2026 Anthropic donated Agent Skills to the Agentic AI Foundation (AAIF) under the Linux Foundation (source).
• A “skill” is a folder containing a SKILL.md that teaches an agent a specific task or domain, plus any scripts or reference material it needs (source).
• Since December 2025, more than 70 agent platforms (including ChatGPT, Cursor, GitHub Copilot, and VS Code) have integrated Agent Skills (source).
• So far, 1 million skills have been published and nearly 280 million installed (source).
• Engineers from Anthropic, Google, OpenAI and others will maintain the standard as individuals; David Soria Parra and Mahesh Murag are Lead Maintainers (source).

What was announced and why it matters for B2B

Anthropic announced on 22 October 2026 that it is donating Agent Skills to the Agentic AI Foundation (AAIF), under the Linux Foundation, to keep the standard open, community-driven and vendor-neutral (source). Since its publication as a standard in December 2025, more than 70 agent platforms—such as ChatGPT, Cursor, GitHub Copilot and VS Code—have integrated Agent Skills (source).

For a B2B website and API estate, this matters because interoperability becomes practical: expose the same actions once and allow multiple agents to execute them safely. That means agents can book demos, query product catalogues, or open and update support tickets—without one-off integrations for each assistant your customers use.

The donation includes the Agent Skills name and logo, the GitHub organisation, and the agentskills.io domain. You can keep using and contributing to the standard without joining the AAIF (source). If you work with Claude, note that ARTIC is a Claude Partner and follows this evolution closely for enterprise rollouts.

What Agent Skills is and how it fits your stack

Agent Skills is an open standard where each “skill” is a folder containing a SKILL.md that teaches an agent a specific task or domain, along with any scripts or reference material needed (source). Anthropic developed skills while working on Claude Code to address gaps in procedural knowledge that many real business tasks require (source).

The community adopted the format rapidly after the December 2025 publication: 1 million skills have been published and nearly 280 million installed so far (source). For technical teams, the skill-as-folder approach makes it straightforward to package task instructions, parameters, examples and helper scripts in a way agents can understand and execute consistently across platforms.

If your goal is to be discoverable and actionable by AI systems, start aligning your content, metadata and endpoints with agent-friendly patterns. This complements broader visibility work such as appearing in mainstream assistants—see our guide on how to appear in AI systems like ChatGPT, Claude or Gemini in 2026.

Implications for booking demos, catalogues and support

With Agent Skills, you can package critical workflows from your website and APIs so AI agents can execute them with clarity and control. Typical B2B scenarios include:

• Demo bookings: availability look-up, lead validation, confirmation and reminders.
• Catalogue queries: filtered search, stock visibility, datasheets and comparisons.
• Support: ticket creation, prioritisation, status updates and documented closure.

The value for go-to-market and customer operations is reach without redundancy. Because more than 70 agent platforms already integrate Agent Skills (source), your packaged actions can be available where users work—whether that’s ChatGPT, Cursor, GitHub Copilot or VS Code—using the same skill. The folder + SKILL.md pattern keeps documentation, scripts and references together for easier distribution and maintenance (source).

If you are exploring dedicated assistants for your processes, see our approach to AI agents for business and how this aligns with the broader enterprise setup of Claude for business.

What to do now: prepare your website and APIs

• Identify 3–5 repeatable, high-impact tasks (e.g., book a demo, query catalogue, open a ticket). Specify inputs, outputs and states.
• Standardise endpoints and contracts: naming, status codes, error shapes and pagination. Include real examples.
• Design permissions and limits: authentication, role-based authorisation, rate limits, quotas and token expiry. Keep an action audit trail.
• Write a clear SKILL.md: purpose, steps, parameters, assumptions, and any scripts or reference material it needs (source).
• Create test environments with synthetic data and edge cases. Ensure idempotency for sensitive operations (e.g., bookings).
• Add observability: structured logs, tracing and metrics per action. Alert on errors, latency and timeouts.
• Version and publish: semantic versioning for the skill, changelogs and automated tests.
• Review compliance and privacy: minimise personal data and define retention windows.

If you plan enterprise-wide deployment and governance around assistants, explore how this fits with Claude for business and technical build-outs via AI development with Claude Code.

Governance, roadmap and security of the standard

Anthropic is donating the Agent Skills name and logo, the GitHub organisation and the agentskills.io domain to the AAIF to ensure neutral governance with community input (source). Engineers from Anthropic, Google, OpenAI and other companies will maintain the standard as individuals, with David Soria Parra and Mahesh Murag as Lead Maintainers (source).

A proposed roadmap will start with versioned releases, a formal proposal process, and security guidance for people who write and install skills (source). Anthropic notes the AAIF—co-founded with Block and OpenAI—also hosts the Model Context Protocol (MCP), reinforcing transparent, community-driven standards for agent interoperability (source).

You can continue using or contributing to Agent Skills without joining the AAIF, and you can follow progress via the standard’s GitHub and community channels referenced from agentskills.io (source).

Frequently asked questions

What is Anthropic’s Agent Skills, in simple terms?

It is an open standard for packaging “skills” that teach an AI agent a specific task or domain. Each skill is a folder with a SKILL.md plus any scripts or reference material the task needs (source).

Does using Agent Skills lock me into one AI assistant?

No. The standard focuses on interoperability. Since December 2025, more than 70 agent platforms have integrated Agent Skills, so the same skill can run across different assistants (source).

Who will maintain the Agent Skills standard going forward?

Anthropic donated the standard to the Agentic AI Foundation (AAIF) under the Linux Foundation. Engineers from multiple companies will maintain it as individuals, with David Soria Parra and Mahesh Murag as Lead Maintainers (source).

Is there guidance on versions and security for skills authors?

Anthropic proposed a roadmap covering versioned releases, a formal proposal process and security guidance for authors and installers of skills (source).

Want to put this to work in your company?

ARTIC is a Claude Partner: we help you roll out Claude across your team, build AI agents connected to your tools and train your people to get value from day one.

Sources

• Donating Agent Skills to the Agentic AI Foundation \ Anthropic

Informative article with the sources linked, up to date on the publication date.