P0
Hard release blocker.
The release does not ship. There is no risk-acceptance path and no override.
Validation — Range
A proving ground, not a promise.
Failure is tested before production finds it.
Range is the mandatory adversarial validation environment. Guard, recovery, tenant isolation, prompt-injection resistance, evidence integrity and Sentinel's own control plane are exercised against controlled scenarios.
What is exercised
The controls that matter when something goes wrong.
Range runs defensive validation against a controlled environment. It confirms that a control behaves as specified under adverse conditions — including conditions where Sentinel itself is the target.
Range is defensive resilience testing against Sentinel’s own environment. It is not an offensive capability and publishes no attack technique.
Validation scenarios
Sample resultsTenant isolation holds under forged workspace claim
Guard
Compromised administrator cannot suppress Watchtower
Watchtower
Sentinel control-plane compromise is contained
Control plane
Evidence integrity survives storage failure
Evidence
AI prompt-injection resistance at the agent boundary
Guard
Recovery completes within continuity objective
Continuity
Queue failure does not drop telemetry
Ingestion
Provider outage degrades without data loss
Continuity
Release gates
Outcomes are stated, not scored.
A scenario holds, degrades, fails, or was not run. Each is tied to a gate that determines whether a release may proceed.
P1
Ships only with documented risk acceptance.
A named owner, a written acceptance and a remediation date. Without all three it is treated as P0.