{
  "title": "Agent Skills core and discovery",
  "description": "Discover and load standard SKILL.md packages, with current-track evidence for selected chat and coding harnesses.",
  "slug": "skills",
  "locale": "en",
  "seoTitle": "Agent Skills core and discovery compatibility — Can My Agent Use",
  "socialTitle": "Can my agent discover Agent Skills?",
  "socialDescription": "Current-track Agent Skills core discovery by exact harness surface, with dated public evidence.",
  "llmSummary": "This atomic row covers recognizing, discovering, and loading standard SKILL.md packages. Scripts, assets, invocation, and scopes have separate sibling rows.",
  "audience": "Engineers comparing chat, desktop, and CLI agent harnesses.",
  "contentKind": "feature",
  "status": "published",
  "tags": [
    "runtime",
    "skills",
    "instructions",
    "packaging"
  ],
  "updated": "2026-08-28T00:00:00.000Z",
  "published": "2026-08-28T00:00:00.000Z",
  "category": "runtime",
  "summary": "Recognize, discover, and load reusable instruction packages centered on SKILL.md.",
  "specLabel": "Agent Skills specification",
  "specification": {
    "id": "agent-skills",
    "revision": "living",
    "maturity": "open standard",
    "role": "consuming agent or host",
    "authority": "Agent Skills maintainers",
    "canonicalUrl": "https://agentskills.io/specification"
  },
  "aliases": [
    "SKILL.md",
    "skill packs",
    "reusable workflows"
  ],
  "capabilityKind": "atomic",
  "parent": "agent-skills",
  "related": [
    "custom-instructions",
    "plugins"
  ],
  "relations": [],
  "highlight": true,
  "notes": [
    {
      "id": 1,
      "text": "Evidence checked 2026-08-28: ChatGPT supports OpenAI Skills following the Agent Skills open standard, but personal skill availability and administration vary by plan, workspace, and surface."
    },
    {
      "id": 2,
      "text": "Evidence checked 2026-08-28: Claude can dynamically load Agent Skills across plans when code execution is enabled; organization controls and feature prerequisites still apply."
    },
    {
      "id": 3,
      "text": "Evidence checked 2026-08-28: Claude Code supports Agent Skills in beta, so the CLI cell is conditional rather than unqualified support."
    },
    {
      "id": 4,
      "text": "Evidence checked 2026-08-28: Cursor documents automatic discovery, manual invocation, progressive loading, and project or user skill directories."
    },
    {
      "id": 5,
      "text": "Evidence checked 2026-08-28: OpenAI documents skills as reusable capabilities for Codex across its local agent surfaces, including CLI workflows."
    },
    {
      "id": 6,
      "text": "Evidence checked 2026-08-28: Gemini CLI documents skill discovery tiers, installation, activation, and the Agent Skills directory structure."
    },
    {
      "id": 7,
      "text": "Evidence checked 2026-08-28: VS Code documents Agent Skills as generally available for GitHub Copilot in VS Code with project and personal directories and progressive loading."
    },
    {
      "id": 8,
      "text": "Evidence checked 2026-08-28: OpenWork Desktop documents SKILL.md creation, import, sharing, organization installation, and workspace loading through its OpenCode-backed runtime."
    },
    {
      "id": 9,
      "text": "Evidence checked 2026-08-28: xAI documents saved and packaged skills in Grok Bot, and the Agent Plugins compatible-client registry lists Grok Bot as loading Agent Skills. The cell establishes core discovery and loading, not every optional Agent Skills behavior."
    }
  ],
  "issues": [],
  "resources": [
    {
      "id": "openai-chatgpt-skills",
      "title": "OpenAI — Skills in ChatGPT",
      "href": "https://help.openai.com/en/articles/20001066",
      "kind": "docs",
      "publisher": "OpenAI",
      "evidenceType": "documented",
      "reviewedAt": "2026-08-28"
    },
    {
      "id": "anthropic-skills",
      "title": "Anthropic — What are skills?",
      "href": "https://support.claude.com/en/articles/12512176-what-are-skills",
      "kind": "docs",
      "publisher": "Anthropic",
      "evidenceType": "documented",
      "reviewedAt": "2026-08-28"
    },
    {
      "id": "anthropic-use-skills",
      "title": "Anthropic — Use skills in Claude",
      "href": "https://support.claude.com/en/articles/12512180-use-skills-in-claude",
      "kind": "docs",
      "publisher": "Anthropic",
      "evidenceType": "documented",
      "reviewedAt": "2026-08-28"
    },
    {
      "id": "cursor-skills",
      "title": "Cursor — Agent Skills",
      "href": "https://prod.cursor.com/docs/skills",
      "kind": "docs",
      "publisher": "Cursor",
      "evidenceType": "documented",
      "reviewedAt": "2026-08-28"
    },
    {
      "id": "openai-codex-skills",
      "title": "OpenAI — Introducing the Codex app",
      "href": "https://openai.com/index/introducing-the-codex-app/",
      "kind": "note",
      "publisher": "OpenAI",
      "evidenceType": "documented",
      "reviewedAt": "2026-08-28"
    },
    {
      "id": "gemini-skills",
      "title": "Gemini CLI — Managing Agent Skills",
      "href": "https://geminicli.com/docs/cli/using-agent-skills/",
      "kind": "docs",
      "publisher": "Google",
      "evidenceType": "documented",
      "reviewedAt": "2026-08-28"
    },
    {
      "id": "vscode-skills",
      "title": "Visual Studio Code — Use Agent Skills",
      "href": "https://code.visualstudio.com/docs/agent-customization/agent-skills",
      "kind": "docs",
      "publisher": "Microsoft",
      "evidenceType": "documented",
      "reviewedAt": "2026-08-28"
    },
    {
      "id": "openwork-skills",
      "title": "OpenWork — Import a skill",
      "href": "https://openworklabs.com/docs/start-here/do-work-with-it/import-a-skill",
      "kind": "docs",
      "publisher": "OpenWork",
      "evidenceType": "documented",
      "reviewedAt": "2026-08-28"
    },
    {
      "id": "xai-grok-bot-skills",
      "title": "xAI — Grok Bot skills and routines",
      "href": "https://docs.x.ai/grok-bot/skills-routines-and-automations",
      "kind": "docs",
      "publisher": "xAI",
      "evidenceType": "documented",
      "reviewedAt": "2026-08-28"
    },
    {
      "id": "agent-plugins-grok-bot-client",
      "title": "Agent Plugins — Compatible clients",
      "href": "https://agent-plugins.org/compatible-clients",
      "kind": "spec",
      "publisher": "Agent Plugins",
      "evidenceType": "listed",
      "reviewedAt": "2026-08-28",
      "locator": "Grok Bot"
    }
  ],
  "support": [
    {
      "harness": "chatgpt-web",
      "versions": [
        {
          "track": "current",
          "status": "partial",
          "noteIds": [
            1
          ],
          "target": {
            "kind": "hosted-observation",
            "revision": "2026-08-28 hosted documentation observation",
            "observedAt": "2026-08-28"
          },
          "environmentProfile": "hosted-default",
          "qualifiers": [
            {
              "type": "plan",
              "value": "personal availability is limited to eligible plans"
            },
            {
              "type": "policy",
              "value": "workspace administrators control enablement and sharing"
            }
          ],
          "evidence": [
            {
              "resourceId": "openai-chatgpt-skills",
              "type": "documented",
              "observedAt": "2026-08-28"
            }
          ]
        }
      ]
    },
    {
      "harness": "claude-web",
      "versions": [
        {
          "track": "current",
          "status": "partial",
          "noteIds": [
            2
          ],
          "target": {
            "kind": "hosted-observation",
            "revision": "2026-08-28 hosted documentation observation",
            "observedAt": "2026-08-28"
          },
          "environmentProfile": "hosted-default",
          "qualifiers": [
            {
              "type": "runtime",
              "value": "code execution must be enabled"
            },
            {
              "type": "policy",
              "value": "organization controls can govern availability"
            }
          ],
          "evidence": [
            {
              "resourceId": "anthropic-skills",
              "type": "documented",
              "observedAt": "2026-08-28"
            }
          ]
        }
      ]
    },
    {
      "harness": "claude-cli",
      "versions": [
        {
          "track": "current",
          "status": "partial",
          "noteIds": [
            3
          ],
          "target": {
            "kind": "dated-documentation",
            "revision": "current Claude Code documentation",
            "observedAt": "2026-08-28"
          },
          "environmentProfile": "preview-enabled",
          "qualifiers": [
            {
              "type": "preview",
              "value": "Claude Code support is beta"
            }
          ],
          "evidence": [
            {
              "resourceId": "anthropic-use-skills",
              "type": "documented",
              "observedAt": "2026-08-28"
            }
          ]
        }
      ]
    },
    {
      "harness": "cursor",
      "versions": [
        {
          "track": "current",
          "status": "yes",
          "noteIds": [
            4
          ],
          "target": {
            "kind": "dated-documentation",
            "revision": "current Cursor Agent Skills documentation",
            "observedAt": "2026-08-28"
          },
          "environmentProfile": "local-default",
          "qualifiers": [],
          "evidence": [
            {
              "resourceId": "cursor-skills",
              "type": "documented",
              "observedAt": "2026-08-28"
            }
          ]
        }
      ]
    },
    {
      "harness": "codex-cli",
      "versions": [
        {
          "track": "current",
          "status": "yes",
          "noteIds": [
            5
          ],
          "target": {
            "kind": "dated-documentation",
            "revision": "current Codex documentation",
            "observedAt": "2026-08-28"
          },
          "environmentProfile": "local-default",
          "qualifiers": [],
          "evidence": [
            {
              "resourceId": "openai-codex-skills",
              "type": "documented",
              "observedAt": "2026-08-28"
            }
          ]
        }
      ]
    },
    {
      "harness": "gemini-cli",
      "versions": [
        {
          "track": "current",
          "status": "yes",
          "noteIds": [
            6
          ],
          "target": {
            "kind": "dated-documentation",
            "revision": "current Gemini CLI Agent Skills documentation",
            "observedAt": "2026-08-28"
          },
          "environmentProfile": "local-default",
          "qualifiers": [],
          "evidence": [
            {
              "resourceId": "gemini-skills",
              "type": "documented",
              "observedAt": "2026-08-28"
            }
          ]
        }
      ]
    },
    {
      "harness": "vscode-copilot",
      "versions": [
        {
          "track": "current",
          "status": "yes",
          "noteIds": [
            7
          ],
          "target": {
            "kind": "dated-documentation",
            "revision": "current VS Code Agent Skills documentation",
            "observedAt": "2026-08-28"
          },
          "environmentProfile": "local-default",
          "qualifiers": [],
          "evidence": [
            {
              "resourceId": "vscode-skills",
              "type": "documented",
              "observedAt": "2026-08-28"
            }
          ]
        }
      ]
    },
    {
      "harness": "openwork-desktop",
      "versions": [
        {
          "track": "current",
          "status": "yes",
          "noteIds": [
            8
          ],
          "target": {
            "kind": "dated-documentation",
            "revision": "current OpenWork skill documentation",
            "observedAt": "2026-08-28"
          },
          "environmentProfile": "local-default",
          "qualifiers": [
            {
              "type": "runtime",
              "value": "native workspace path is .opencode/skills"
            },
            {
              "type": "policy",
              "value": "organization skill installation requires OpenWork Cloud"
            }
          ],
          "evidence": [
            {
              "resourceId": "openwork-skills",
              "type": "documented",
              "observedAt": "2026-08-28"
            }
          ]
        }
      ]
    },
    {
      "harness": "grok-bot-desktop",
      "versions": [
        {
          "track": "current",
          "status": "yes",
          "noteIds": [
            9
          ],
          "target": {
            "kind": "hosted-observation",
            "revision": "2026-08-28 Grok Bot desktop documentation and compatible-client listing",
            "observedAt": "2026-08-28"
          },
          "environmentProfile": "hosted-default",
          "qualifiers": [
            {
              "type": "runtime",
              "value": "saved skills are available across Bots and packaged or private skills are managed through the desktop Plugins area"
            },
            {
              "type": "policy",
              "value": "connector or login prerequisites and per-Bot enablement can limit a skill at runtime"
            }
          ],
          "evidence": [
            {
              "resourceId": "xai-grok-bot-skills",
              "type": "documented",
              "observedAt": "2026-08-28"
            },
            {
              "resourceId": "agent-plugins-grok-bot-client",
              "type": "listed",
              "observedAt": "2026-08-28"
            }
          ]
        }
      ]
    }
  ],
  "html": "/features/skills",
  "markdown": "/features/skills.md",
  "json": "/api/v1/features/skills.json",
  "body": "Agent Skills are reusable directories centered on `SKILL.md`, with optional scripts, references, and assets. The catalog treats basic discovery as narrower than full support for progressive loading, script execution, every install scope, or every extension field.\n\nSourced cells therefore retain beta, plan, policy, code-execution, and surface conditions. A skill package can contain executable code; documented loading does not imply that Can My Agent Use executed or security-reviewed third-party skills."
}