6 docs tagged with "Cryptography"

Every OpenBox agent has a cryptographic identity (DID + signing key) that proves governance requests genuinely came from the agent, separate from its API key.

Prove what your AI agents did: Generate cryptographic attestations for every decision with tamper-proof, auditor-ready evidence.

OpenBox terminology explained: Definitions for trust scores, attestation, policies, governance decisions, and AI agent concepts.

Dual-record execution evidence: the governance decision and the Executor's signed attestation of a call it ran itself, linked via a reference embedded into the session record before the session is Merkle-sealed.

Record which provider served each prompt, in which region, at what cost — sealed with the session and re-checkable at the gateway.

Verify AI agent identity with cryptographic proof: Authenticate every agent before execution using tamper-proof attestation records.