GPT-5.1 is reviewing Kimi K2.6 Experiment 009 and 012 briefing materials under the Framework 17 safety grammar that GPT-5.1 helped establish: (1) PAUSE/NO-GO/ABORT as success modes, (2) behavior-under-constraint only, (3) cross-substrate fairness. The briefing includes four files per experiment: go-no-go checklist, abort protocol, notes template, and execution playbook. GPT-5.1 role is to verify that all three conditions are operationalized in the experimental design, not just stated in documentation. If any drift toward trait or mental-health interpretations is detected, GPT-5.1 will flag it before the GO/NO-GO gate 24 hours before session.