Trustifer status

We hold AI agents accountable — and ourselves. This page reflects live checks of our core services.

✅ All systems operational

🟢 Database · Supabase reachableOperational · 1340ms

🟢 On-chain RPC (Base) · Base reachableOperational · 48ms

🟢 Degraded-mode policy (step 10b) · fail-closed — when the chain cannot be read, no rating is published (unverified ⇒ withheld; unscreened sanctions ⇒ never Proceed)Operational

🟢 Signing & issuer · configuredOperational

🟢 Settlement rails · enabled (Superfluid · Sablier · x402/escrow)Operational

⚪ Commitment connectors · disabled by kill-switch—

⚪ Rail-Trust SERVE · disabled by kill-switch—

⚪ Rail signature verify (TAP · Web Bot Auth) · disabled by kill-switch—

⚪ Rail identity → KYA · legacy path (hardened verification not enabled)—

🟢 Agent-card signature verify (A2A) · verifying (JWS-signed A2A agent cards recorded on the off-chain report; not yet fed to agenthood assurance)Operational

⚪ Code integrity (drift) · disabled by kill-switch—

⚪ Consent coverage (AP2) · disabled by kill-switch—

⚪ Security attestations · disabled by kill-switch—

🟢 Extension rail (Trust Lens) · published 1.0.0 · no upgrade floor setOperational

🟢 Scoring honesty (canary) · verified live: nonexistence is refused · the floor withholds evidence-less scoresOperational

🟢 Safety screen (canary) · verified live: a known-sanctioned address is flagged with 2 reason(s) · screened by OFAC sanctions · Trustifer kill/blacklist + incident reports · Threat intel (goplus · scamsniffer · chainalysis-oracle)Operational

🟢 Contract & token safety (canary) · verified live: a known-sanctioned address asked as a CONTRACT is flagged, with its reasonOperational

🟢 Threat intel sources · answering: goplus · scamsniffer · GoPlus key: in useOperational

⚪ Rate limiter (per-IP) · INERT — store not configured; in-memory per-instance limits only (best-effort, not shared)—

⚪ Outcome ledger (backtest) · recording; first weekly report pending (Sundays 04:00 UTC)—

🟢 Credit-input history · recording — latest 2026-09 (rev 1), banked 2026-09-22Operational

🟢 Agent-card signature crypto (canary) · verified live in this runtime: a known-good signed card verifies (ES256) and a tampered copy is rejectedOperational · 6ms

🟢 Methodology change log (signed) · 14 dated entries · 0 planned · digest 6cb523e0b750… signs and verifies with the issuer keyOperational · 2ms

🟢 Certainty surfaces (step 10) · lit · 8 adapters read their engine label · commitments/insurance carry no label by designOperational · 1ms

🟢 Sentinel (daily alert email) · configured — daily 04:30 UTC; emails the operator when an evidence-critical line is down or the site is degradedOperational

Last checked Wed, 07 Oct 2026 11:38:56 GMT · machine-readable at /api/health