{
  "title": "Identity vs path: two AI-crawler log tests on one log (Unsourced, October 2026)",
  "site": "unsourced.app",
  "window": "2026-10-01 to 2026-10-28 (UTC)",
  "license": "CC BY 4.0. Credit: Unsourced (unsourced.app). Path-test method: VantaWeb (vantaweb.io/blog/ai-crawler-traffic-audit/).",
  "visits": 6805,
  "distinct_ips": 739,
  "distinct_paths": 1012,
  "identity_test": "Claimed AI crawler checked against the operator's currently published IP ranges and forward-confirmed reverse DNS. verified = proven; unverified = positive contradiction (IP outside the published range, or forward-confirmed rDNS owned by someone else); unknown = no published footprint to check against.",
  "path_test": "published = path in any version of the site's sitemaps, any app route, the standard machine files, or a directory the pages themselves load (/assets, /og, /samples, /api/js, /demo). Everything else = never_published.",
  "cross": [
    {"identity": "verified", "path": "published", "visits": 2480},
    {"identity": "verified", "path": "never_published", "visits": 22},
    {"identity": "unverified", "path": "published", "visits": 342},
    {"identity": "unverified", "path": "never_published", "visits": 1705},
    {"identity": "unknown", "path": "published", "visits": 2075},
    {"identity": "unknown", "path": "never_published", "visits": 181}
  ],
  "scanner_campaign": {"visits": 1578, "servers": 9, "network": "Google Cloud customer VMs (googleusercontent.com)", "behaviour": "each server rotated 9 to 10 AI-crawler names and requested secret-file paths such as /.env and cloud credential files", "dates": "3 to 28 October 2026"},
  "limits": "One site, four weeks, a site about AI crawlers. Identity needs current IP lists, so only recent visits are used. No claim about who sent any request."
}