Language‑model agents are increasingly embedded in operational workflows that involve approvals, bindings, and payments, creating commitments. An approval that is valid when issued may cease to authorize execution once relevant facts, evidence, policies, or authorities change—this is the evaluation‑execution gap. We introduce a state‑bound runtime assurance layer that separates probabilistic measurement from deterministic adjudication, binds an execution allowance to the decision‑relevant basis of an approval, invalidates that allowance after logged changes, and rechecks admissibility at dispatch. This distinguishes the continued validity of a prior approval from a current‑state policy check. We outline eight failure modes, an assurance model, a reference architecture, and prospective interventions with retrospective remediation. In a scripted 42‑case replay, the state‑bound configuration identified 15 approvals that later became invalid—approvals a deliberately non‑re‑evaluating point‑in‑time comparator would not revisit; it held 14 affected actions before dispatch and compensated one landed effect. These results illustrate behavior under the stipulated policies and event orders, not superiority over dispatch‑time enforcement designs or production effectiveness. Under explicit interception, ingestion, dependency, and serialization assumptions, the model establishes dispatch‑time validity within the registered execution surface.
Review