The receipt is the durable proof.
It binds boundary status, owner context, commitment ID, and signature material in one machine-readable record.
Craton adds an independently verifiable receipt layer before irreversible actions execute. Before execution, Craton returns a signed record that is verifiable offline, without relying on Craton's runtime. Your host keeps execution control; Craton supplies the pre-execution proof layer.
Before the action moves, Craton returns signed evidence your team can store with the downstream record and verify without calling Craton back.
It binds boundary status, owner context, commitment ID, and signature material in one machine-readable record.
Show what was evaluated, which boundary applied, and which commitment was signed before the action proceeded.
Verify with Ed25519, the public JWKS, and the published Craton protocol. No callback required.
Craton returns signed evidence; your host system decides how to apply its policy.
Review the protocol, verifier, schema, JWKS, and offline verifier kit from public resources.
One risky action becomes one boundary check, one signed receipt, and one audit-ready proof reference.
The path is scoped to a specific high-risk action, system, owner, and threshold.
The host calls Craton before the action moves and receives the boundary result.
The receipt keeps the commitment and verification material together.
Anyone with the receipt, schema, and public key can verify the evidence later.
Start with the action. Craton maps it to a signed boundary path and asks only for the facts needed to issue verifiable receipt evidence.
Describe only the action you want covered. Craton maps the boundary and receipt evidence, not your full business payload. The path can be prepared from the scope facts you provide here.
Use one concrete action with the trigger, amount, or review gap if you know it.
Craton recommends the boundary pattern automatically. These options are available for teams that want to inspect or override the mapping.
Before it executes, the boundary is already fixed.
The same boundary result survives every handoff.
Exceptions can happen, but they must leave proof.
A completed action leaves independently verifiable evidence.
When a high-risk action has a clear trigger, Craton fixes the boundary before execution and returns a machine-verifiable boundary result. Whatever the host does next, the signed receipt can be verified and carried forward.
Start with one high-risk action. Craton turns it into a machine-verifiable boundary path, then provisions API access after signing and payment are confirmed. Broader rollout can be added later through continuation or expansion documentation.
Contract, readiness packet, protocol spec, schema, and standard public keys.
Send partner, gate, system, decision type, and owner before the action moves.
The response includes verdict, receipt, commitment ID, and request ID.
Use the public protocol and keys to verify the receipt payload.
Pass commitment_id as the downstream reference.
Save the path first. Continuation, expansion, and end-state choices stay available after the initial scope is clear.
{}
Start with one machine-verifiable path. Add new gates, systems, or organizations only after the first runtime receipt exists.
Add another action under the same gate when it can share the same receipt shape.
Decision identity, threshold checks, manual override, and risk escalation. Each gate adds a new boundary surface without changing the primitive.
Carry the recorded result from one internal system into another after the runtime receipt exists.
When multiple organizations depend on the same runtime boundary receipt, Craton becomes shared infrastructure.
One sensitive operation. One insertion point. One production-bound path that returns a machine-verifiable result when your system calls Craton.