Choose the action families your agent supports. CommandLayer capabilities define action semantics, schemas, discovery surfaces, and receipt verification paths.
Choose the action families your agent supports. CommandLayer capabilities define action semantics, schemas, discovery surfaces, and receipt verification paths.
A capability is not just a verb. It can include request schemas, receipt schemas, examples, agent cards, ENS records, MCP tool definitions, OpenAPI operations, SDK usage, and verifier support.
Capability semantics define verb meaning and expected execution intent.
Request schemas define valid inputs before runtime execution begins.
Runtime executes the action and signs receipts using signer identity and key material.
Canonical receipts include metadata.proof fields for JSON canonicalization, SHA-256 hashing, and Ed25519 signatures.
MCP bridge and API discovery surfaces expose callable tools and operations.
Verifier validation confirms proof integrity, and tamper invalidation fails altered receipts.
Trust Verification v1 is live in production. The remaining capability families are namespace package / future family tracks for expansion.
Select capability groups for your agent stack. Export local manifests, SDK config, agent cards, ENS records, and discovery drafts.