Picture this. A developer asks an AI assistant to "scale down staging to save costs." The AI, helpful as always, executes: kubectl scale deployment critical-api --replicas=0 -n production. Wrong namespace. Right outcome, wrong cluster. The API is down.