Agent developer
Stop holding keys, stop hand-rolling policy. Get a signed audit trail for free. Start at the quickstart — daemon up + curl + signed receipt in < 5 minutes.
Agent developer
Stop holding keys, stop hand-rolling policy. Get a signed audit trail for free. Start at the quickstart — daemon up + curl + signed receipt in < 5 minutes.
Compliance / security
Cryptographically verifiable proof every agent action was authorised, with offline re-derivation. Read audit log and capsule v2.
Sponsor team
Receive a signed APRP envelope upstream of execution. Replay-safe, policy-checked, audit-linked. Read sponsor adapters.
Auditor / judge
Download one JSON file, run one verifier command, reconstruct what was authorised. Read APRP wire format.
SBO3L is the cryptographically verifiable trust layer for autonomous AI agents. Every action your agent takes — pay, swap, store, compute, coordinate — passes through SBO3L’s policy boundary first: schema validation, deterministic policy decision, multi-scope budgeting, Ed25519-signed receipts, hash-chained audit log, sponsor adapter routing.
Output: a self-contained Passport capsule anyone can verify offline against the agent’s published Ed25519 pubkey alone — no daemon, no network, no RPC.
Phase 3 extends the trust boundary across protocols, chains, and tenants. New conceptual guides:
Audit anchoring
Merkle-root commits to L1 turn the local audit chain into ledger-verifiable proof. On-chain audit anchoring →
Multi-tenant
One daemon, N tenants, hard isolation across DB / signer / HTTP layers (V010 migration). Multi-tenant →
Marketplace
Content-addressed signed policies; reputation scored from real fleet observations. Marketplace →
Token-gated identity
Bind agent authority to ERC-721 / 1155 ownership; AnyOf / AllOf composites; time windows. Token-gated identity →
Cross-protocol composition
14 framework adapters share one audit chain — wallet vs mandate, end-to-end. Cross-protocol composition →