A practical glossary for Sentinel SCA concepts used in AI governance, autonomous systems, execution-boundary enforcement, audit integrity, and replay verification.
The determination that a proposed autonomous action satisfies governance requirements before execution.
The final control point where a proposed action becomes a real action in software, infrastructure, physical systems, APIs, or autonomous environments.
An operational layer that evaluates proposed actions while the system is running and returns ADMIT, REVIEW, or DENY decisions.
A record proving that governance evaluation occurred and showing the decision, evidence, identity, capability, or enforcement status associated with an action.
The practice of limiting agents, devices, or services to explicit operational capabilities rather than granting broad execution authority.
A safeguard that prevents old or duplicate requests from being reused to trigger repeated or unauthorized execution.
A tamper-evident sequence of governance records used to preserve decision history and support forensic verification.
The ability to reconstruct or review prior governance decisions using preserved evidence and audit records.
A safety model where uncertainty, invalid evidence, expired approval, or missing authority results in non-execution.
A governance workflow where higher-risk actions are halted until an authorized human approves or rejects execution.
A limited time period during which an approval or authorization remains valid before revalidation is required.
An enforcement response returned by an execution boundary indicating whether an action was halted, executed, rejected, expired, or otherwise handled.