PTERI's threat model — what's protected and what isn't

PTERI treats security as a property of architecture, not a layer bolted on afterwards.

Threats eliminated by design

• Credential theft and replay attacks — there are no shared secrets to steal
• Phishing of passwords or OTP codes — there are none
• Impersonation by infrastructure — the platform can verify but never sign on your behalf
• AI forgery — proofs require a private key that never leaves your control

Explicitly out of scope

PTERI does not protect against:

• Compromised devices
• User-approved malware
• Physical coercion

Keep your device secure and review every approval request before you sign it.