TOKEN PROFILE
ES256 with a distinct MSE+JWT protected type
The terminal receipt is a compact ES256 JWT. Its protected header must contain typ=MSE+JWT, alg=ES256 and a retained MandateShield JWKS kid. Its exact string aud binds the PROCESSOR audience that caused the transition. It is evidence of a past state transition and can never authorize another provider submission.
TERMINAL STATE
Action, origin, destination and outcome remain inseparable
The signed claims bind the authorization, original decision receipt, optional permit, transition, input, attempt, protocol, provider, payee, amount and resource. They record exactly one action and its legal terminal state:
COMMIT → COMMITTEDfor a recorded successful result, including a late canonical success after a priorSETTLEMENT_UNKNOWN.RELEASE → RELEASEDfor a confirmed failure or non-submission before the release boundary.EXPIRE → EXPIREDfor an unused reserved authorization derived from the MandateShield ledger.SETTLE_UNKNOWN → SETTLEMENT_UNKNOWNwith outcomeSETTLEMENT_UNKNOWNwhen submission ambiguity remains conservatively charged.
A nullable provider, resource, payment reference or occurrence time is an explicit absence, never an invitation for a verifier to infer missing facts.
EVIDENCE CLASSES
Provenance and caller-independent checking are separate fields
| Class | Meaning |
|---|---|
CALLER_ASSERTED | The authenticated PROCESSOR adapter supplied the result; it is not provider-signed evidence. |
FACILITATOR_SIGNED | A facilitator-signed artifact is bound, but it was checked without trusting the caller only when independent_verification is true. |
PROVIDER_API_VERIFIED | An exact provider API result is bound. independent_verification is true only when MandateShield performed the configured lookup without trusting the caller report. |
CHAIN_FINALIZED | A canonical chain-finality result is bound. independent_verification is true only when MandateShield performed the configured RPC checks without trusting the caller report. |
UNRESOLVED | Submission may have occurred but no conclusive terminal provider or chain result is available. |
LEDGER_DERIVED | The terminal fact follows only from MandateShield's own authorization ledger, such as unused reservation expiry. |
A direct manual COMMIT or RELEASE still emits CALLER_ASSERTED for its authenticated processor_result and sets independent_verification=false. Signed Stripe webhooks are also caller hints, not independent evidence. Unused expiry remains LEDGER_DERIVED, while an inconclusive provider or chain observation remains UNRESOLVED.
Hosted provider reconciliation may emit PROVIDER_API_VERIFIED with independent_verification=true after an exact Stripe API lookup, or CHAIN_FINALIZED with that boolean after canonical x402 chain verification. Only supported, non-conflicting evidence can autonomously COMMIT or RELEASE and produce that receipt. Pending, unknown and conflicting observations remain fail-closed. A stronger evidence_class never overrides the boolean by itself. Here, “independent” means the configured check did not trust the caller report; it does not mean an independent organization, external audit or certification.
VERIFICATION BOUNDARY
Cryptographic authenticity is precise but deliberately bounded
Offline verification can prove that MandateShield signed these exact terminal claims with the identified retained key. It does not independently prove the underlying provider assertion, settlement, chain finality, facilitator policy or customer deployment. Consumers must evaluate evidence_class, evidence_digest and independent_verification together.
This is vendor-published integration infrastructure, not an independent audit, provider certification or statement that a provider currently mandates MandateShield.