claim.check
Claim support and contradiction. Execution availability depends on authentication, authorization, provider readiness, and production entitlements.
Switch between an adequate and inadequate evidence state to inspect the expected gate behavior.
Claim supported within authorized scope. Continue only if every independent execution check also passes.
status: verified
next: execution policy check
receipt: inspectableClaim support and contradiction. Execution availability depends on authentication, authorization, provider readiness, and production entitlements.
Verify the cited proposition against its source. Execution availability depends on authentication, authorization, provider readiness, and production entitlements.
Pre-action evidence and authority checks. Execution availability depends on authentication, authorization, provider readiness, and production entitlements.
Validate proposed transactional scope. Execution availability depends on authentication, authorization, provider readiness, and production entitlements.
Inspect what was checked and its limits. Execution availability depends on authentication, authorization, provider readiness, and production entitlements.
Track defined evidence states and change. Execution availability depends on authentication, authorization, provider readiness, and production entitlements.
// Illustrative integration pattern — not a live executable SDK example
const intent = agent.proposedAction();
const result = await riv.verify({
action: intent,
sources: context.sources,
policy: authority.scope,
spendLimit: budget.limit
});
if (result.status !== "verified") return stop(result.limitations);
await agent.execute(intent);
await riv.receipt(result);An agent first defines intent, evidence, policy and spend boundary. The verification result informs whether to act, stop, or request review. A receipt preserves the reviewed state.
This snippet illustrates the architecture. It is not a claim that the specific JavaScript methods shown are available as a production SDK.
Read integration guide ↗R/IV registers 29 machine capabilities: 23 released one-time operations and six recurring monitors in limited availability. Payment is operation- and rail-specific. Inspect the live status contract before integration; a catalog entry never grants authority to pay or execute.
The x402 / Base USDC rail returns a live payment challenge for each released one-time operation. Capability state, inputs, limits and receipt behavior are defined by the live registry.
Fifteen operations meeting the current Card / Link floor are available through MPP. The other eight released one-time operations are x402-only. Recurring monitors are review-gated and are not automated machine subscriptions.
Operation-level release status does not establish independent platform certification, benchmark superiority, universal delivery quality, or authorization beyond the exact request.
Public documentation covers callable interfaces, prerequisites, response boundaries and receipts. Proprietary implementation, credentials, private orchestration, provider controls and unpublished research remain restricted.
Use the live catalog first, inspect a no-charge preflight, authorize the exact capability and spend boundary, execute through the protected endpoint, then validate and preserve the signed receipt.
GET /api/v1/preflight?capability_id=citation.verify. Authenticated quotes use POST /api/v1/quote. A quote is not a purchase.
OAuth authorization metadata ↗. Bind capability, input scope and maximum authorized cost before execution.
For machine-paid work, use the released https://pay.robertduraniv.com/x402/execute/{capability_id} route or the MPP route returned by its live catalog. Follow the issued challenge; do not construct headers or infer an unsupported path.
POST /api/v1/receipts/verify. Preserve the receipt, evidence verdict, action gate, limitations and execution identity with the downstream action.