Contract
Request/result schemas, service version, execution mode, acceptance conditions, and evidence requirements define what a buyer is purchasing.
CommandLayer keeps a small set of open, inspectable rules for canonicalization, signatures, receipt scopes, schemas, and verification. The commercial factory uses those rules to manufacture paid machine outcomes.
Request/result schemas, service version, execution mode, acceptance conditions, and evidence requirements define what a buyer is purchasing.
Providers, primitive graphs, retries, state, budgets, and routing can change behind a stable contract.
Canonical hashes, signatures, signer identity, timestamps, and scoped evidence make execution independently inspectable.
The same covered payload must produce the same canonical representation.
deterministicThe verifier checks whether the declared signer signed the covered receipt payload.
integrityExecution proof covers execution fields. Settlement proof, when present, remains a separate scope.
separatedA service may prove that declared checks passed without claiming the output is objectively true in every external sense.
contractualA service can be published through different naming, registry, discovery, and settlement systems without forcing those systems into the core receipt contract.
ENS, HTTPS identity, wallets, ERC-8004 registration, and other identifiers can coexist.
Owned catalog, OpenAPI, MCP registries, x402 discovery, 8004scan, and future directories are publication surfaces.
Credits, cards, x402, invoicing, or future rails can settle a quote while metering remains canonical.
A payment can prove that value settled. An execution receipt can prove what service ran, what inputs/outputs were hashed, what checks were applied, and who signed the result. Neither should impersonate the other.
{
"service": "example@1.0",
"execution": {
"input_hash": "sha256:...",
"output_hash": "sha256:...",
"accepted": true
},
"proofs": ["execution", "optional settlement scope"]
}