Skip to content

X-VR-001 Verification Result Contract

The Verification Result Contract describes a verifier's report for presented evidence under a recipient's policy. It records what was checked and what remains uncertain. A technical ALLOW is not an institution's acceptance decision.

The current design is 3.0.0 proposal.4, a proposal rather than a released interoperable contract. The recipient controls its policy, trust inputs and acceptance decision. Only that institution may record L7_acceptance; other producers leave it NOT_ASSESSED.

Inputs and outputs

Inputs are the presentation, current time, recipient-held policy profile identified by id, version and digest, and named trust inputs. A producer-supplied policy is not authority for the recipient's requirements.

The result has seven layers with the following exact keys. The floor checks are required by the candidate contract independently of additional recipient requirements.

Layer key Concern Floor checks
L1_parse Serialisation, binding version and credential type parse
L2_signature Signature, issuer identity and key control; applicable presentation binding signature, issuer_identity
L3_authority Issuer authority under the recipient's accepted trust inputs issuer_authority
L4_status Validity period and status validity_period, status
L5_semantics Asset binding, addressee, engagement and conflicts between sources asset_binding
L6_policy Purpose and required disclosure under recipient policy purpose
L7_acceptance The recipient institution's recorded decision None

Checks use satisfied, rejected, unknown, unavailable, stale, not_checked, not_applicable or unsupported. Outcomes for L1_parse through L6_policy are PASS, FAIL, UNKNOWN, UNSUPPORTED, NOT_APPLICABLE or NOT_ASSESSED. The technical projection is ALLOW only when all six layers PASS; otherwise it is DENY. Unknown evidence never becomes PASS.

L7_acceptance has separate outcomes: NOT_ASSESSED or DECISION_RECORDED. A recorded decision is accepted_for_scope, declined or referred_for_manual_review. Only the recipient institution may record that decision. Describing this vocabulary does not change the proposal's adoption state.

Disclosure diagnostics record requested_paths, withheld_required_paths and over_disclosed_paths, containing paths only. If a required observation date is withheld, the disclosure check is unavailable, the affected layer is UNKNOWN and the projection is DENY. Keep this insufficient-evidence outcome distinct from a contradiction, rejected evidence or invalid signature.

Scope and authority

A result is limited to its presented subset, policy and time of verification. It expires with the shortest validity or freshness limit it depends on. Credential signatures, issuer authority, status, asset binding and disclosure requirements are separate checks. Recognition is scoped to the recipient's held policy and does not imply equivalence between frameworks or jurisdictions.

For example, a lender's verifier might establish all six technical layers for a valuation, including its asset binding, authority and currentness. It reports ALLOW while acceptance remains NOT_ASSESSED until the lender records its decision.

Wire identifiers and draft limits

The current proposal uses internal-placeholder:unregistered-namespace:schema:verification-result:3.0.0-proposal.4; this is not a registered identifier.

Prototype schema and reference-checker tests address result consistency only. They do not demonstrate credential signature verification or a complete verifier. Enforcement of recipient-controlled asset binding and status freshness remains incomplete in the assessed reference implementations. Schema validation alone does not enforce all cross-field rules. Recipient-supplied asset references, recipient-bounded freshness limits and independent adversarial verification remain necessary. See Verifier behaviour and Assurance and release for the distinction between requirements and demonstrated implementation behaviour.