Agent-facing governance files

Machine-Readable Governance

Verifies /llms.txt, /llms.json, /.well-known/ai-agent.json, agent-start, AI summary, and safe-read-order delivery for bounded machine readers. This page is read-only and does not certify, approve, train, probe, or mutate anything.

LLMS / agent-start / safe read order

Machine-Readable Governance

Machine-Readable Governance is a read-only orientation and direct-file delivery layer. It does not certify the site, mutate DNS, validate credentials, train models, probe private networks, approve drafts, close incidents, or claim exact glyph translation.

Package: 1.29.0
Seed: 2026.06.11.031
Surface: machine-readable-governance

Copyable JSON evidence

{
    "schemaVersion": "1.0.0",
    "packageVersion": "1.29.0",
    "seededVersion": "2026.06.11.031",
    "surface": "machine-readable-governance",
    "generatedAtUtc": "2026-09-17T22:08:38Z",
    "liveFinding": "The v1.20 live site returned 404 for /llms.txt while the uploaded strategy requires LLMS and agent-start governance files to orient machine readers before summaries.",
    "deliveryModes": [
        {
            "path": "/llms.txt",
            "format": "text/plain",
            "purpose": "first human/agent orientation file"
        },
        {
            "path": "/llms.json",
            "format": "application/json",
            "purpose": "structured LLMS governance payload"
        },
        {
            "path": "/.well-known/ai-agent.json",
            "format": "application/json",
            "purpose": "well-known agent boundary manifest"
        },
        {
            "path": "/agent-start.json",
            "format": "application/json",
            "purpose": "agent-start orientation packet"
        },
        {
            "path": "/ai-summary.json",
            "format": "application/json",
            "purpose": "bounded AI-readable summary"
        },
        {
            "path": "/safe-read-order.json",
            "format": "application/json",
            "purpose": "ordered route list for reading boundaries before summaries"
        },
        {
            "path": "/barkbowl-machine-readable-governance.json",
            "format": "application/json",
            "purpose": "low-dependency direct governance beacon"
        }
    ],
    "operatorChecks": [
        "Open /llms.txt and confirm it returns 200 text, package 1.29.0, seed 2026.06.11.031, no runtime-authority claim, and a safe read order.",
        "Open /llms.json, /.well-known/ai-agent.json, /agent-start.json, /ai-summary.json, and /safe-read-order.json and confirm bounded JSON.",
        "Run Runtime Verification after seeding and confirm Machine-Readable Governance and LLMS Governance route shells exist.",
        "Confirm static package files exist under root, theme, and plugin packages for offline review."
    ],
    "claimBoundary": "Machine-Readable Governance is a read-only orientation and direct-file delivery layer. It does not certify the site, mutate DNS, validate credentials, train models, probe private networks, approve drafts, close incidents, or claim exact glyph translation.",
    "humanReviewRequired": true
}