Audit any AI-capability paper against an editorial reporting checklist: which model variant was tested, when the evaluation ran, how completely the model was elicited, and whether the conclusion stayed scoped to those tested models. Companion tool to Gringras and Salahshoor, 2026.