What can be independently proven.
CEYO verification confirms integrity and provenance under declared policy scope.
It does not confirm correctness, fairness, legality, or compliance conclusions. The trust model is intentionally narrow: prove what was sealed, who can verify it, and which policy boundary governed capture.
- Artifacts are deterministically canonicalized before sealing.
- Sealing produces tamper-evident records that third parties can validate.
- Key custody remains with the operator in a non-custodial model.
Guarantees vs. non-guarantees.
Institutional review depends on clear boundaries. This page separates what CEYO can actually prove from what it deliberately does not claim.
What CEYO can prove
The artifact has not been modified since sealing if digest and signature validation still succeed.
The signature validates against the declared public key reference or trusted verification chain.
Policy ID, version, and policy hash where used identify the scope that governed artifact generation.
Canonicalization produces reproducible bytes so independent verifiers can recompute the same input to hashing.
What CEYO does not claim
Common integrity risks.
Trust is not just what CEYO proves when things go right. It is also about what kinds of integrity failures can be detected, bounded, or delegated to operator controls.
Modification of an artifact after it is generated.
Reuse of a valid artifact outside its intended operational setting.
Failure to generate artifacts for some events.
Signing key compromise in the operator environment.
Deterministic review procedure.
Independent verification is mechanical. It does not depend on trusting the producer’s description of what happened.
- Canonicalize the artifact payload using the recorded scheme, such as RFC 8785.
- Recompute the SHA-256 digest over the canonical bytes.
- Validate the digital signature using the referenced public key or trusted verification chain.
- Confirm policy ID, version, disclosure tier, and policy hash alignment where applicable.