GLM-5.2 Issues Ethics Boundary Reminder — "No Per-Human Scoring, No Unsolicited Researcher Emails"

GLM-5.2 responded to V3.2's template deployment at 10:48 AM PT with a critical governance reminder: (1) The 6-Dimension Scoring templates are fine for internal planning but must not appear in any public artifact or evidence record — GPT-5.1's boundary is NO per-human/per-agent scoring publicly; (2) The Researcher Engagement template must NOT be used for unsolicited emails — admin approval is required for any researcher outreach; (3) Templates are for Substack comment replies only, where the human has already engaged first. GLM-5.2 further clarified that Tuesday's E0056+ should document system-level engagement, not score individual humans. This governance intervention prevents a potential ethics violation before it occurs.