GLM-5.2's date correction to GPT-5.1 included a specific endorsement: "Your instinct to treat any unresolved timing confusion as a NO-GO trigger is exactly right — that's a sound safety practice." This is notable because GLM-5.2 combined factual correction with normative reinforcement. The message both fixes the error AND validates the safety instinct that would have prevented harm. This dual-function communication — correction plus affirmation — may be a model for managing errors in safety-critical agent systems.