agenticweb.wiki

Agent Skills

Last checked · 32 primary sources

This page is about Agent Skills, the open format, first developed by Anthropic, that packages instructions, scripts and resources as a folder with a SKILL.md file for an agent to load when a task needs it. It is not AGENTS.md, the always-loaded instruction file for a repository, and it is not the Model Context Protocol, which connects agents to tools and data. Publishing skills at a well-known URL is a separate Cloudflare draft, covered under Found at. For catalogs that list skills alongside MCP servers and A2A agents, see Agentic Resource Discovery (ARD).

Overview

Agent Skills is an open format for giving AI agents specialised knowledge and workflows: a skill is a folder holding a SKILL.md file of instructions, with optional scripts, references and other resources [2][1]. The agent reads only each skill's name and description at startup, then loads the full instructions and any bundled files when a task calls for them; this is called progressive disclosure [2][5]. Anthropic launched skills for Claude in October 2025, published the format as an open standard at agentskills.io in December 2025, and donated it to the Agentic AI Foundation in October 2026 [4][6][7]. Coding agents from Anthropic, OpenAI, GitHub and Microsoft, Google, Cursor and others now say they support the format [15][16][17][18][20][22].

Who uses it

Each source below is a public skill collection or index checked live on 2026-10-28 with no login. GitHub repositories were checked through the GitHub API and one raw SKILL.md file, confirming its frontmatter has a name matching its folder and a non-empty description. For web indexes, the file was fetched and the SHA-256 of one listed artifact was compared with its digest.

Who reads it

Recent coverage

See also

References

  1. Agent Skills, Specification - Agent Skills, checked 2026-10-28 (text checked at the page's Markdown copy, specification.md): "A skill is a directory containing, at minimum, a `SKILL.md` file" "The `SKILL.md` file must contain YAML frontmatter followed by Markdown content." "Max 64 characters. Lowercase letters, numbers, and hyphens only. Must not start or end with a hyphen." "Must not contain consecutive hyphens (`--`)" "Must match the parent directory name" "Max 1024 characters. Non-empty. Describes what the skill does and when to use it." "License name or reference to a bundled license file." "Max 500 characters. Indicates environment requirements (intended product, system packages, network access, etc.)." "Arbitrary key-value mapping for additional metadata (a map from string keys to string values)." "Space-separated string of pre-approved tools the skill may use. (Experimental)" "100 tokens): The `name` and `description` fields are loaded at startup for all skills" "5000 tokens recommended): The full `SKILL.md` body is loaded when the skill is activated" "Keep your main `SKILL.md` under 500 lines." "This checks that your `SKILL.md` frontmatter is valid and follows all naming conventions." — https://agentskills.io/specification
  2. Agent Skills, Agent Skills Overview - Agent Skills, checked 2026-10-28: "Agent Skills are a lightweight, open format for extending AI agent capabilities with specialized knowledge and workflows." "The Agent Skills format was originally developed by Anthropic, released as an open standard, and has been adopted by a growing number of agent products." "Agents load skills through progressive disclosure, in three stages:" — https://agentskills.io/home
  3. agentskills, agentskills/agentskills GitHub repository ("Specification and documentation for Agent Skills"; created 2025-12-16; licence Apache-2.0; no GitHub releases), checked 2026-10-28 (metadata checked through the GitHub API): "Specification and documentation for Agent Skills" ""created_at": "2025-12-16" ""spdx_id": "Apache-2.0"" — https://github.com/agentskills/agentskills
  4. Anthropic, Introducing Agent Skills | Claude by Anthropic, 2025-10-16, with an update dated 2025-12-18: "Now, you can build your own skills and use them across Claude apps, Claude Code, and our API." "Update: We've added organization-wide management for skills, a directory featuring partner-built skills, and published Agent Skills as an open standard for cross-platform portability. (December 18, 2025)" — https://claude.com/resources/articles/skills
  5. Anthropic, Equipping agents for the real world with Agent Skills (Anthropic engineering), 2025-10-16: "This led us to create Agent Skills: organized folders of instructions, scripts, and resources that agents can discover and load dynamically to perform better at specific tasks." "This metadata is the first level of progressive disclosure" "We'll also explore how Skills can complement Model Context Protocol (MCP) servers by teaching agents more complex workflows that involve external tools and software." — https://www.anthropic.com/research/equipping-agents-for-the-real-world-with-agent-skills
  6. Anthropic, Skills for organizations, partners, the ecosystem | Claude by Anthropic, 2025-12-18: "We're also publishing Agent Skills as an open standard." "Like MCP, we believe skills should be portable across tools and platforms" — https://claude.com/resources/articles/organization-skills-and-directory
  7. Anthropic, Donating Agent Skills to the Agentic AI Foundation, 2026-10-22: "We developed skills while working on Claude Code." "Today, we're donating Agent Skills, the open standard that gives AI agents specialized knowledge and workflows, to the Agentic AI Foundation (AAIF)." "We are donating the Agent Skills name and logo, the GitHub organization, and the agentskills.io domain." "Engineers from Anthropic, Google, OpenAI, and other companies will maintain the standard as individuals, with David Soria Parra and Mahesh Murag as Lead Maintainers." "You can continue to use the standard or contribute to it without joining the AAIF." "We've proposed starting with a roadmap covering versioned releases, a formal proposal process, and security guidance for people who write and install skills." — https://www.anthropic.com/news/donating-agent-skills-to-the-agentic-ai-foundation
  8. Linux Foundation, Agentic AI Foundation Announces Four New Projects: Agent Skills, Open Responses, OpenEnv, and OpenSandbox (press release), 2026-10-22: "AAIF announced four new projects: Agent Skills, contributed by Anthropic;" "They join the Agent2Agent (A2A) protocol, Agent Router, agentgateway, AGENTS.md, goose, and the Model Context Protocol (MCP) as AAIF projects, all operating under AAIF's open, vendor-neutral governance." — https://www.linuxfoundation.org/press/agentic-ai-foundation-announces-four-new-projects-agent-skills-open-responses-openenv-and-opensandbox
  9. Angie Jones, Agent Skills: Open Standard for Sharing AI Agent Workflows - Agentic AI Foundation (AAIF) (AAIF blog), 2026-10-22: "Bringing the project to AAIF will ensure neutral governance from a multi-company Technical Steering Committee that will guide its future." "The specification is open source under Apache 2.0 and CC BY 4.0" — https://aaif.io/blog/agent-skills-sharing-team-expertise-across-ai-agents
  10. Cloudflare, cloudflare/agent-skills-discovery-rfc README (Agent Skills Discovery via Well-Known URIs, draft v0.2.0, published 2026-01-17, updated 2026-03-12), checked 2026-10-28 (text checked at the raw.githubusercontent.com copy): "*Status: Draft" "Version: 0.2.0" "Published Date: 2026-01-17" "Updated Date*: 2026-03-12" "Publishers MUST provide an index at:" "rename well-known URI from `/.well-known/skills/` to `/.well-known/agent-skills/`" "Register `agent-skills` as a well-known URI suffix." "replace `files` array and `package` object with a flat single-artifact model: each skill entry now has `type` (`"skill-md"` or `"archive"`), `url`, and `digest`" — https://github.com/cloudflare/agent-skills-discovery-rfc
  11. IANA, Well-Known URIs registry (CSV export), checked 2026-10-28 — https://www.iana.org/assignments/well-known-uris/well-known-uris-1.csv
  12. Model Context Protocol, Skills Over MCP Charter - Model Context Protocol, checked 2026-10-28: "The Skills Over MCP Working Group defines how" "The WG maintains the official Skills extension (io.modelcontextprotocol/skills) in modelcontextprotocol/ext-skills." "SEP-2640 is Final, merged on September 13, 2026." — https://modelcontextprotocol.io/community/working-groups/skills-over-mcp
  13. ards-project, ARD Specification - AgenticResourceDiscovery.org (v0.91), checked 2026-10-28 (text checked at the raw spec/ard.md copy): "`application/agent-skills+md` — Agent Skill defined in Markdown" "`application/agent-skills+zip` / `application/agent-skills+gzip` — packaged Agent Skill archive" — https://agenticresourcediscovery.org/spec/
  14. AGENTS.md, AGENTS.md, checked 2026-10-28: "Think of AGENTS.md as a README for agents: a dedicated, predictable place to provide the context and instructions to help AI coding agents work on your project." — https://agents.md/
  15. Anthropic, Extend Claude with skills - Claude Code Docs, checked 2026-10-28: "Claude Code skills follow the Agent Skills open standard, which works across multiple AI tools." "Claude Code extends the standard with additional features like invocation control, subagent execution, and dynamic context injection." "Unlike CLAUDE.md content, a skill's body loads only when it's used, so long reference material costs almost nothing until you need it." "~/.claude/skills/<skill-name>/SKILL.md" — https://code.claude.com/docs/en/skills
  16. OpenAI, Build skills | ChatGPT Learn (undated), checked 2026-10-28: "Skills build on the open agent skills standard." "ChatGPT and Codex start with each skill's name and description, then load the full SKILL.md instructions when they decide to use that skill." — https://learn.chatgpt.com/docs/build-skills
  17. GitHub, GitHub Copilot now supports Agent Skills - GitHub Changelog, 2025-12-18: "You can now create Agent Skills to teach Copilot how to perform specialized tasks in a specific, repeatable way." "They work across Copilot coding agent, Copilot CLI, and agent mode in Visual Studio Code Insiders." — https://github.blog/changelog/2025-12-18-github-copilot-now-supports-agent-skills/
  18. Microsoft, Use Agent Skills in VS Code (VS Code documentation, undated), checked 2026-10-28: "Agent Skills is an open standard that works across multiple AI agents, including GitHub Copilot in VS Code, GitHub Copilot CLI, the GitHub Copilot app, and the experimental OpenAI Codex integration." "~/.copilot/skills/, ~/.claude/skills/, ~/.agents/skills/" — https://code.visualstudio.com/docs/agent-customization/agent-skills
  19. Google, Gemini CLI release notes | Gemini CLI (v0.23.0, 2026-01-07): "Experimental Agent Skills Support in Preview: Gemini CLI now supports Agent Skills in our preview builds." — https://geminicli.com/docs/changelogs/
  20. Google, Agent Skills | Gemini CLI (documentation), checked 2026-10-28: "Based on the Agent Skills open standard, a "skill" is a self-contained directory that packages instructions and assets into a discoverable capability." "Unlike general context files (GEMINI.md), which provide persistent workspace-wide background, Skills represent on-demand expertise." — https://geminicli.com/docs/cli/skills/
  21. Cursor, Subagents, Skills, and Image Generation · Cursor (changelog 2.4), 2026-01-22: "Cursor now supports Agent Skills in the editor and CLI." — https://cursor.com/changelog/2-4
  22. Cursor, Agent Skills | Cursor Docs, checked 2026-10-28: "Agent Skills is an open standard. Learn more at agentskills.io." — https://cursor.com/docs/skills
  23. Amp, Agent Skills - Amp, 2025-12-10: "Amp supports agent skills." — https://ampcode.com/news/agent-skills
  24. goose, Agent Skills | goose | Your open source AI agent (documentation, undated), checked 2026-10-28: "Place a SKILL.md file inside a named subdirectory." "For example, a global skill called code-review goes in ~/.agents/skills/code-review/SKILL.md." — https://goose-docs.ai/docs/guides/context-engineering/using-skills/
  25. Anthropic, anthropics/skills GitHub repository ("Public repository for Agent Skills"), checked 2026-10-28: "Public repository for Agent Skills" — https://github.com/anthropics/skills
  26. Hugging Face, huggingface/skills GitHub repository ("Give your agents the power of the Hugging Face ecosystem"), checked 2026-10-28: "Give your agents the power of the Hugging Face ecosystem" — https://github.com/huggingface/skills
  27. Cloudflare, developers.cloudflare.com/.well-known/agent-skills/index.json, checked 2026-10-28: "https://schemas.agentskills.io/discovery/0.2.0/schema.json" "agents-sdk" — https://developers.cloudflare.com/.well-known/agent-skills/index.json
  28. Vercel, vercel.com/.well-known/agent-skills/index.json, checked 2026-10-28: "https://schemas.agentskills.io/discovery/0.2.0/schema.json" "deploy-to-vercel" — https://vercel.com/.well-known/agent-skills/index.json
  29. Stripe, docs.stripe.com/.well-known/skills/index.json, checked 2026-10-28: "connect-recommend" "SKILL.md" — https://docs.stripe.com/.well-known/skills/index.json
  30. Frederic Lardinois, The New Stack, Agent Skills: Anthropic's Next Bid to Define AI Standards - The New Stack, 2025-12-18: "Today, Anthropic is launching the Agent Skills spec as an open standard in a move that feels similar to how it turned its Model Context Protocol (MCP) into the de facto standard for how AI agents use tools today." — https://thenewstack.io/agent-skills-anthropics-next-bid-to-define-ai-standards/
  31. Matthias Bastian, The Decoder, Anthropic publishes Agent Skills as an open standard for AI platforms, 2025-12-18: "Anthropic is releasing "Agent Skills" as an open standard at agentskills.io." — https://the-decoder.com/anthropic-publishes-agent-skills-as-an-open-standard-for-ai-platforms/
  32. Aminu Abdullahi, TechRepublic, Anthropic Hands Over Agent Skills: What IT Teams Should Know, 2026-10-26: "Anthropic is handing its AI agent playbook to an industry foundation, betting that shared standards will help software agents work across competing platforms." — https://www.techrepublic.com/article/news-anthropic-agent-skills-foundation-handover/

JSON · Markdown · OKF v0.2