Real-time domain trust scoring for the Cloudflare One browser extension — closing the 21-hour phishing gap that blocklists can't reach.
Cloudflare One secures two layers today. The browser is the third — and it's where work happens.
Extension collects signals. Edge scores them. Gateway enforces policy.
Hooks webNavigation API at every domain change. Collects domain name, scans rendered DOM for credential forms, detects BITB patterns.
Manifest V3 Chrome/EdgeHybrid analysis: 6-signal heuristic engine (typosquatting, entropy, TLD risk) plus Workers AI semantic classification. Cached globally via KV.
Workers Workers AI KVTrust score feeds into admin-configured rules. The extension provides the signal — the admin's policy decides: allow, warn, isolate, or block.
Gateway RBIType any domain below to see the scoring engine in action.