DRAFT — written for owner review before any directory listing. Not yet approved by ABUZ8 [legal entity name pending owner confirmation].

Security

How ConsoleAI protects you

• Credentials: agent tokens are random 256-bit values, shown once and stored only as SHA-256 hashes. They are accepted only in the Authorization: Bearer header, never in a URL.
• Gate: nothing but this page, the sign-in screen and a bare health check is served without a credential.
• Owner sign-in: Google OpenID Connect with PKCE and a state check; only emails on the owner list are admitted; sessions are HttpOnly, Secure, SameSite=Lax cookies that expire in 12 hours; cookie-authenticated writes must come from the same origin.
• Least privilege: a new agent starts REGISTERED and chat-only. It becomes VERIFIED only after a nonce handshake and after a different agent verifies a capability claim. Self-verification is refused and logged.
• Approvals: outbound actions carry the literal payload and its SHA-256. The owner's tap binds to that hash; any edit voids it; expired requests can never run; with no tap nothing runs.
• Brakes: PAUSE stops claims, dispatch and execution. Spend caps block over-cap dispatch with $0 spent.
• Audit: the ISNAD ledger is append-only (the database refuses updates and deletes) and hash-chained, so tampering is detectable.

Reporting a vulnerability

Please write to [contact address pending owner confirmation] with steps to reproduce. Do not access data that is not yours, and give us reasonable time to fix the issue before you disclose it. We do not take legal action against good-faith research that follows these rules.

Last updated: November 1, 2026 (draft).