Skip to content

Operator guide

You're an operator if you're the person standing up, configuring, and running a Yutha swarm. That includes:

  • Deploying the control plane and choosing a storage backend.
  • Choosing the swarm's topology — closed, open, or hybrid.
  • Authoring and activating the constitution that governs the swarm.
  • Managing operator credentials, including revocation.
  • Anchoring the receipt log on Sui for third-party verifiability (optional).
  • Monitoring receipts and responding to enforcement events.

If you're building agents that join a swarm someone else operates, you're a developer — see the developer guide instead.

Start here

  • Quickstart — the 30-minute initiator path. Stand up a control plane, activate a constitution, register an operator credential, send a first envelope, observe receipts.
  • Authoring constitutions — how to write Cedar+ policy that says what you mean.
  • Operator credentials — how the operator identity works, how to rotate it, how to revoke an agent.
  • Sui anchoring — opt-in cryptographic verifiability via on-chain Merkle commitments.
  • Signer backends — overview — opt-in key custody: hold the control plane's Ed25519 signing key in HashiCorp Vault, GCP KMS, or Azure Managed HSM instead of process memory. Links to the per-backend runbooks.
  • Attestor backends — overview — opt-in admission attestation: verify every Register against an external workload-identity system (SPIFFE/SPIRE or OIDC). Links to the per-backend runbooks.
  • Enterprise identity end-to-end — the integrated playbook that combines a Signer backend with an Attestor backend in one production deployment.
  • Monitoring & receipts — what to watch, what to alert on.
  • Deployment — Postgres backend, scaling, single-tenant defaults.