Machine-Verifiable Identity
Machine-verifiable identity binds an agent, service, or workload identity to cryptographic evidence about the environment executing it. Instead of accepting a name or public key on its own, a verifier can require proof of the hardware, software measurement, and policy behind that identity.
A common pattern is to place a public key in caller-controlled fields of an attestation quote. After verification, messages signed by that key can be attributed to the attested workload.
Machine identity is different from human identity. It answers questions such as:
- Which workload controls this key?
- Which TEE and software measurement produced the evidence?
- Does the platform meet the required security policy?
- Is the evidence fresh enough for this interaction?
This model supports verifiable agents and services whose reputation or permissions depend on hardware-backed execution.