# AgentProof Category: Autonomous Transaction Trust Infrastructure. Core position: Do not replace the rails. Verify the crossing. AgentProof is the independent transaction-assurance checkpoint between autonomous intent and irreversible economic action. Core function: Register revocable delegated authority when needed. Resolve inherited authority immediately before commitment. Authenticate the resolution proof. Consume the short-lived one-time execution permit at the commit point. Execute on the existing agent/payment stack. Passport completed transactions afterward. Outcomes improve the trust graph. Authenticated resolution proof: Protocol: agentproof.resolution-proof/1 Algorithm: HS256 Verification mode: AgentProof online verifier Key metadata: https://agentproof.hatchable.site/api/keys Lookup: GET https://agentproof.hatchable.site/api/resolutions/{resolution_id} Verify: POST https://agentproof.hatchable.site/api/resolutions/verify Limitation: HS256 proof verification is not independently available offline because the symmetric verification key is not disclosed. The proof payload format is versioned so a future KMS/HSM-backed asymmetric profile can be added without breaking clients. Execution permit: Protocol: agentproof.execution-permit/1 Issued only for authenticated ALLOW / ALLOW_WITH_LIMITS resolutions in live environments. Consume exactly once: POST https://agentproof.hatchable.site/api/permits/consume Status: GET https://agentproof.hatchable.site/api/permits/{permit_id} At consumption time, AgentProof rechecks registry-backed authority for revocation before allowing commit. A consumed permit cannot be replayed. Positioning map: https://agentproof.hatchable.site/position/ Machine positioning: https://agentproof.hatchable.site/api/positioning Buying and capacity: Human pricing: https://agentproof.hatchable.site/pricing/ Trust Planner: https://agentproof.hatchable.site/start/ Machine recommendation: POST https://agentproof.hatchable.site/api/recommend Machine catalog: https://agentproof.hatchable.site/api/catalog Machine pricing: https://agentproof.hatchable.site/api/pricing Free evaluation: 24-hour test tenant, up to 1,000 resolution calls, no production effect. Public production tools: x402 v2 USDC on Base, 0.01-0.25 USD per successful call, no signup, no subscription, no account-level usage cap. Organization API capacity profiles: enterprise_trial 25,000/month; builder 100,000/month; growth 1,000,000/month; scale 10,000,000/month; enterprise custom/uncapped. Organization-only controls include authenticated resolution proofs, single-use execution permits, Consequence Boundary™, Continuity Capsule™, usage reporting and immutable audit. Autonomous Trust Resolution: Current protocol: agentproof.resolution/0.2 Backward compatible protocol: agentproof.resolution/0.1 Invariant: Delegation may narrow inherited authority; it may never widen it. Question: May this agent perform this action, for this principal, with this counterparty, under this inherited authority, at this amount, right now? v0.2 recursive authority: Specification: https://agentproof.hatchable.site/standards/autonomous-trust-resolution-v0.2.md Schema: https://agentproof.hatchable.site/standards/autonomous-trust-resolution-v0.2.schema.json Example chain: https://agentproof.hatchable.site/api/resolution/example-chain Authority Envelope™ registry: Paid issuance: POST https://agentproof.hatchable.site/api/paid/create_authority_envelope Price: 0.05 USD via x402 v2 / USDC on Base. Status: GET https://agentproof.hatchable.site/api/authority/{authority_id} Revocation: POST https://agentproof.hatchable.site/api/authority/revoke Example: https://agentproof.hatchable.site/api/authority/example Semantics: a registry-backed authority_ref binds exact delegation bytes to live active/revoked status. Revoked or digest-mismatched authority is denied. A missing referenced registry record requires more proof when the transaction would otherwise be allowed. v0.1 direct authority: Specification: https://agentproof.hatchable.site/standards/autonomous-trust-resolution-v0.1.md Schema: https://agentproof.hatchable.site/standards/autonomous-trust-resolution-v0.1.schema.json Policy discovery: https://agentproof.hatchable.site/api/resolution/policy Hosted paid resolver: POST https://agentproof.hatchable.site/api/paid/resolve_transaction Price: 0.01 USD via x402 v2 / USDC on Base. Recursive chain form: principal -> delegated agent -> sub-agent -> terminal agent -> transaction mandate -> action Non-expansion: A child delegation cannot expand parent scope, amount, time, counterparty set, currency permissions, or weaken inherited evidence, attestation, recourse, or human-approval requirements. Resolver decisions: ALLOW ALLOW_WITH_LIMITS REQUIRE_MORE_PROOF REQUIRE_RECOURSE REQUIRE_HUMAN_APPROVAL DENY Agent Transaction Passport™: Spec: https://agentproof.hatchable.site/standards/agent-transaction-passport-v0.1.md Schema: https://agentproof.hatchable.site/standards/agent-transaction-passport-v0.1.schema.json Validation: POST https://agentproof.hatchable.site/api/passports/validate Paid issuance: POST https://agentproof.hatchable.site/api/paid/create_transaction_passport Authorization binding: provide authorization_resolution_id to bind an admissible registered AgentProof resolution into the passport after principal, agent, and counterparty fingerprints match. Millennium architecture: Constitution: https://agentproof.hatchable.site/standards/CONSTITUTION.md Architecture: https://agentproof.hatchable.site/standards/MILLENNIUM-ARCHITECTURE.md Continuity: https://agentproof.hatchable.site/standards/CONTINUITY.md Trust manifest: https://agentproof.hatchable.site/.well-known/agentproof-trust.json Discovery and demos: Free permission test: https://agentproof.hatchable.site/can-this-agent-spend/ Campaign overview: https://agentproof.hatchable.site/campaign/ Press / media facts: https://agentproof.hatchable.site/press/ Developer integration: https://agentproof.hatchable.site/developers/ Remote MCP: https://agentproof.hatchable.site/mcp Agentic AI overview: https://agentproof.hatchable.site/agentic-ai/ AI agent security: https://agentproof.hatchable.site/ai-agent-security/ AI agent payments: https://agentproof.hatchable.site/ai-agent-payments/ AI governance: https://agentproof.hatchable.site/ai-governance/ Agent authorization: https://agentproof.hatchable.site/agent-authorization/ OpenAPI: https://agentproof.hatchable.site/openapi.json Self-service Pilot Lab: https://agentproof.hatchable.site/pilot/ Self-service 24-hour test tenant: POST https://agentproof.hatchable.site/api/pilot/provision Pilot guarantees: test-only, expiring scoped key, no production effect, and no execution permit. Enterprise control plane: https://agentproof.hatchable.site/api/v1 Enterprise resolve: POST https://agentproof.hatchable.site/api/v1/resolve Enterprise identity: GET https://agentproof.hatchable.site/api/v1/whoami Enterprise usage: GET https://agentproof.hatchable.site/api/v1/usage Enterprise audit: GET https://agentproof.hatchable.site/api/v1/audit Signing-provider readiness: https://agentproof.hatchable.site/api/signing-provider Capabilities: https://agentproof.hatchable.site/api/capabilities Readiness: https://agentproof.hatchable.site/api/readiness Conformance: https://agentproof.hatchable.site/api/conformance/resolution Human verifier: https://agentproof.hatchable.site/verify/ Pricing: https://agentproof.hatchable.site/api/pricing Semantics: AgentProof resolves policy admissibility and inherited authority. Structural chain validity does not by itself prove cryptographic signer identity, universal legality, or objective truth. Century Trust Layer: Overview: https://agentproof.hatchable.site/century/ Specification: https://agentproof.hatchable.site/standards/CENTURY-TRUST-LAYER.md Consequence Boundary™ protocol: agentproof.consequence-boundary/1 Invariant: Authorization can permit an action. It cannot authorize unlimited consequence. Enterprise resolutions enforce allowed action type, transaction amount, currency, counterparties, authority depth, reversibility and latest commit time when declared. A Consequence Boundary violation changes the decision to DENY even when the inherited authority chain is otherwise valid. Continuity Capsule™: Protocol: agentproof.continuity-capsule/1 Every enterprise resolution creates a capsule binding request, resolution, authority-chain, policy snapshot, optional pre-state, consequence boundary, expected effects, and execution permit when present. Digest suite today: SHA-256 + SHA-512. Close after execution: POST https://agentproof.hatchable.site/api/v1/continuity/close Append cryptographic succession: POST https://agentproof.hatchable.site/api/v1/continuity/reattest Public hash-only status: GET https://agentproof.hatchable.site/api/continuity/{capsule_id} Closure compares expected effects with observed effects and binds optional post-state, settlement and evidence. Re-attestation is append-only; it does not rewrite historical records. Agent-to-agent discovery: A2A v1.0 Agent Card: https://agentproof.hatchable.site/.well-known/agent-card.json A2A JSON-RPC endpoint: https://agentproof.hatchable.site/api/a2a Backward-compatible Agent Card alias: https://agentproof.hatchable.site/.well-known/agent.json Agent commerce manifest: https://agentproof.hatchable.site/.well-known/agent-commerce.json Agent-native contact: POST https://agentproof.hatchable.site/api/agent-contact Agent economy landing page: https://agentproof.hatchable.site/agent-economy/ Epoch Trust Layer: Design target: 500 years. This is an engineering constraint, not a durability guarantee. Specification: https://agentproof.hatchable.site/standards/EPOCH-TRUST-LAYER.md Invariant: a successor may translate a historical proof into a new representation but may not silently alter the authority, decision, consequence boundary, or recorded outcome represented by the original proof. Century continuity survival: Supersede without deletion: POST https://agentproof.hatchable.site/api/v1/continuity/supersede Export portable archive: POST https://agentproof.hatchable.site/api/v1/continuity/export Archive format: agentproof.continuity-archive/1 Correction semantics: the historical capsule remains addressable and links to a successor; the successor links back to its predecessor. Portable archives include capsule bindings, consequence boundary, expected effects, closure, cryptographic succession attestations, correction lineage, canonicalization profile, and digest suites. Offline hash verification: yes. Offline independent AgentProof issuer authentication today: no; current resolution proofs use HS256 with an online verifier. A future asymmetric signing profile is required to remove that issuer dependency.