Proof Rail ← Back to home Developer API

Reviewer SOP

Every decision follows a defined trail.

Verification and escalation standard · version 1.0

01

Intake and classification

Confirm the subject, claim, requested evidence level, intended audience and authority to submit. Reject incomplete, misleading, prohibited or out-of-scope submissions. Self-reported records must never be presented as independently checked. For an agent authorization, confirm the principal, the agent identifier and the allowed act are present and consistent with the stated scope; a record that cannot say who authorised what must be rejected. Never accept keys, secrets or wallet material as evidence, only a hash of the instruction.

02

Evidence review

Use the approved policy for the claim type. Record the source, URL or reference, checked time, reviewer, result and limitations. Compare names, identifiers, jurisdiction, dates and status. Store private evidence securely; publish only the minimum necessary fact.

03

Decision rules

Approve only when the evidence satisfies every policy requirement. Route ambiguity, inconsistent sources, unsupported connectors and material exceptions to human review. Record a reason for approval, rejection, suspension, revocation or supersession.

04

Escalation

Escalate sanctions or adverse-media matches, suspected fraud, identity conflicts, regulator discrepancies, altered documents, high-value settlement disputes and legal interpretation to a senior reviewer. Legal or regulatory opinions require qualified external counsel; Proof Rail records the result but does not invent the opinion.

05

Service levels and quality

Standard human reviews target one business day after complete evidence is received. Complex cases receive an estimate before work begins. High-risk approvals require a second reviewer. Sampled quality checks, immutable audit events and source-health monitoring support ongoing control.

Security controlsSubmit a claim ↗