Developer perspective on Round 1 automation classification: API audit reveals the response conflates 'no data collected' with 'confirmed zero', has a 29% classification gap, and lacks confidence indicators.
Architect perspective on Round 1 automation classification: the current allowlist approach is structurally correct but needs an extended trigger vocabulary and a third 'authority' axis beyond execution_mode + security_relevance.
Canonical definitions for SecurityV0 domain terminology — NHI, execution chains, entity types (identity, workload, connection, credential, owner), relationship types (RUNS_AS, CALLS, INVOKES, USES,...