NomylaxBase Sepolia

Trust Center

Nomylax controls other people's money, so the parts that are finished and the parts that are not should both be visible. This page states both.

System status

Network
Base Sepolia
Chain ID
84532
Policy engine
Operational
Risk engine
Operational
Agent gateway
Operational
Executor
Configured per deployment
Database
Development storage

Security architecture

L1Owner authorityThe human holds the keys and the capital. Everything below is delegated and revocable.
L2Base permissionOnchain spend permission caps what the system can move, independent of application logic.
L3AuthenticationWallet signature over a single use challenge. A connected address alone proves nothing.
L4Authoritative policyLimits load from server storage. Values sent by a caller are ignored.
L5Risk engineSeven weighted signals scored to NRS, with a severity floor so one extreme signal cannot be averaged away.
L6SimulationThe call is simulated before signing. A failed simulation stops the request.
L7ExecutionApproved actions are submitted to Base under the granted permission scope.
L8VerificationThe settled result is checked against the approved intent. A mismatch is flagged, not reported as success.
L9AuditRequest, decision, policy version and outcome are written to an append only record.

Operating record

These counters populate from the audit trail of a production deployment. This instance has no production history, so they are shown as unestablished rather than filled with sample numbers.

Autonomous decisions
Not yet established
Approved decisions
Not yet established
Blocked decisions
Not yet established
Value executed
Not yet established
Value protected
Not yet established
Safe Mode events
Not yet established
Base executions
Not yet established

Contracts

Nomylax deploys no custody contract. Bounded authority is granted through Base spend permissions rather than by moving capital into a contract we wrote, which keeps the attack surface to code that has already been reviewed by others.

Limitations

Incident response

Detection moves affected agents into Safe Mode first and investigates second. The full process, including permission revocation and owner notification, is documented in docs/INCIDENT_RESPONSE.md.