Your agent didn't hallucinate; it exceeded its authority

Content filters block unsafe output but can't tell whether an agent was authorized to issue a refund, touch a production system, or commit the company to an external action — different problems most enterprises only solve the first.