{
  "$schema": "https://github.com/cloudflare/agent-skills-discovery-rfc",
  "generatedAt": "2026-06-30T18:00:00Z",
  "skills": [
    {
      "name": "Understand Agent Action Governance",
      "type": "markdown",
      "description": "Understand the Linkality category, decision labels, and routed-action boundary.",
      "url": "https://linkality.com/.well-known/agent-skills/understand-agent-action-governance.md",
      "sha256": "2f44324381505eebf9275ed0c51ba1f19ee74a520cf024c64614d0be7932a1c6",
      "safety": "Use this skill to explain Agent Action Governance without claiming universal coverage. State that Linkality governs actions routed through MCP/API. Point users to policy, approval, audit, and receipt pages for details.",
      "docs": "https://linkality.com/agent-action-governance"
    },
    {
      "name": "Evaluate MCP Action Risk",
      "type": "markdown",
      "description": "Classify risk for a proposed MCP tool action before it proceeds.",
      "url": "https://linkality.com/.well-known/agent-skills/evaluate-mcp-action-risk.md",
      "sha256": "1ddc699aae6b7cd7bfb03c66a63ce42203d6581a50f155444cc4107ee83ae20a",
      "safety": "Use simulation for uncertain or high-risk actions. Treat destructive, production, credential, database, financial, email, Slack, webhook, and deploy-like actions as governed. Do not proceed on BLOCKED or pending approval results.",
      "docs": "https://linkality.com/ai-agent-risk-map"
    },
    {
      "name": "Request Human Approval",
      "type": "markdown",
      "description": "Explain how to wait for a real human decision on APPROVAL_REQUIRED actions.",
      "url": "https://linkality.com/.well-known/agent-skills/request-human-approval.md",
      "sha256": "f02d320901cc1969a2e708fe298b2175c40837a7b27f902596d746706be8664b",
      "safety": "Never fabricate approval. Poll linkality_get_decision or direct the reviewer to the Linkality app. Approval applies to the exact recorded action only.",
      "docs": "https://linkality.com/ai-agent-approval-workflows"
    },
    {
      "name": "Verify Signed Receipt",
      "type": "markdown",
      "description": "Verify hash and signature posture for a Linkality receipt.",
      "url": "https://linkality.com/.well-known/agent-skills/verify-signed-receipt.md",
      "sha256": "4b53b1ceccc0edf1c507b57be4b6d78552ea207b2716a5245c63eb901842ba4a",
      "safety": "A hash match proves integrity. A signature proves origin only if it validates against a trusted public key. Unsigned receipts are integrity-only.",
      "docs": "https://linkality.com/signed-receipts-for-ai-agents"
    },
    {
      "name": "Query Governance Ledger",
      "type": "markdown",
      "description": "Inspect ledger entries and receipt history without exposing private payloads.",
      "url": "https://linkality.com/.well-known/agent-skills/query-governance-ledger.md",
      "sha256": "0c8d7cc19bc4323a671983bead524caeeb90498dac85b6fa61ad6f46ae3d52f2",
      "safety": "Use ledger reads for audit and status only. Do not expose raw secrets, headers, tokens, or private payloads. Respect workspace scope.",
      "docs": "https://linkality.com/ai-agent-audit-logs"
    },
    {
      "name": "Inspect Policy Pack",
      "type": "markdown",
      "description": "Read active policy posture before proposing an action.",
      "url": "https://linkality.com/.well-known/agent-skills/inspect-policy-pack.md",
      "sha256": "b037b12c879fefceae7a119f257476a9f6f1c0b4ee569e1fac87d9fd3cc29e8d",
      "safety": "Call linkality_get_policies first when deciding what is governed. Treat policy results as authoritative for routed actions. Do not infer permissions from missing public docs.",
      "docs": "https://linkality.com/ai-agent-policy-engine"
    },
    {
      "name": "Understand GitHub Broker Governance",
      "type": "markdown",
      "description": "Describe GitHub governance as narrow, approval-gated, and certification-gated unless enabled.",
      "url": "https://linkality.com/.well-known/agent-skills/understand-github-broker-governance.md",
      "sha256": "7301162eacdc40093b6bab03d58c550d0cc01b0be66c5d9d2796ad926c47f426",
      "safety": "Do not claim unrestricted GitHub writes. Do not claim automatic merges or protected-branch bypass. Agents never receive standing GitHub credentials from public discovery.",
      "docs": "https://linkality.com/github-agent-governance"
    },
    {
      "name": "Understand Database Broker Read-Only Mode",
      "type": "markdown",
      "description": "Describe safe database governance posture without exposing private schemas.",
      "url": "https://linkality.com/.well-known/agent-skills/understand-database-broker-read-only-mode.md",
      "sha256": "4ea8653fe730fd00e8a392303d5189577878afe36f1501e825e679945f25fef8",
      "safety": "Treat public database broker posture as bounded read-only. Approval does not enable writes. Do not expose private schemas, connection strings, or customer data.",
      "docs": "https://linkality.com/database-agent-governance"
    }
  ]
}