Constitutional identity of this entity (human, agent, or device). Verified, not trusted — anchored to a declared DID / provenance.
Professional: identity graph, signed provenance, revocation status.
Activation is a constitutional capability, not a one-time event: verify → repair → resume → trust. Never re-births the entity; never creates a second identity (L1).
Natural-language interaction. Every turn is selected through USE when a capability choice exists; otherwise it is a direct constitutional action.
Sovereign, content-addressed knowledge the entity may draw on. Read-only by default; provenance preserved.
Declarative tasks the entity executes under constitutional selection. Each is recorded (USE-4) before execution.
Optional capability adapters — never core deps (UAM). Selectable through USE; capability-free path always preferred (USE-5).
Sovereign state and the USE audit trail (MEMORY-LD). Auditable, append-only, content-addressed.
Constitutional defaults. Changes are governed actions, recorded and reversible.