Engagement that informs ZDG-AIS, ZDG-AFW, and ZDG-ACP placement.
Start by mapping the workflow, not just the model. Identify execution families, approval thresholds, operator handoffs, and evidence gaps before rollout.
ZDG-AFW — Agent Firewall.
Evaluate actions before execution. Allow, block, and approval are explicit runtime states, not implied policy intentions.
ZDG-BB / ZDG-FR — Black Box and Flight Recorder.
Preserve the governed chain so teams can inspect what happened later. Replay-linked proof turns runtime control into something that can be reviewed and defended.
ZDG-AIS — Agent Integrity Signals.
Add bounded containment when agent behavior drifts toward risky patterns. The goal is controlled intervention, not theatrical shutdowns.
ZDG-ACP — Approval Control Plane.
Give operators a real surface for approvals, exceptions, proceed actions, and evidence handoff. This is where policy meets accountable operation.
Review the stakeholder proof story first, then open the raw proof and platform pages for deeper inspection.