Status: Accepted
Date: 2026-07-22
Decider: Chelsea Kelly-Reif
A credible operational exercise needs an empty isolated target, read-back, evidence matrices, and five operational probes. No lawful native export, source profile, or real target has passed discovery. A generic adapter or simulated restoration would freeze an untested abstraction and invite an unsupported success claim.
Add one strict, synthetic-only exercise-plan contract. It may validate:
- a separately captured baseline and declared per-dimension coverage;
- a customer-obtainable synthetic export mechanism;
- an empty, isolated, egress-blocked target with automations disabled;
- prohibition of production data;
- target read-back and raw evidence requirements; and
- exactly five lookup, relationship, attachment, allow, and deny probes.
The validator contains no URL, credential, command, mapping, load, read-back, or
result interface. Its CLI status is synthetic_protocol_valid with decision
scope plan_only_no_target_execution.
- Build a generic target protocol: rejected until one real source-target pair tests the abstraction.
- Build a simulated target: rejected because a simulator can only prove behavior it invents.
- Keep the protocol prose-only: rejected because omitted sandbox and evidence controls are cheaply detectable before a real exercise.
- Teams can prepare a complete safety/evidence checklist offline.
- A valid plan makes no claim that declared controls exist.
- Source profiles, target interfaces, read-back, and result assurance remain blocked on discovery and a real sandbox.