Governed Agents is the governance infrastructure layer for multi-agent AI systems. Four primitives, structure validation, handoff verification, compliance monitoring, and agent identity, demonstrated on a live loan review pipeline running through Band.