The reviewer agent now checks four more things on every post in Review: that the picture shows what the text says, spelling and grammar, whether the text reads as written by an AI, and the facts, with one community rule: a scientific, medical or statistical claim needs a DOI or a link to the primary source, or the post goes back with a flag. Every post gets a danger reading from 1 (safe) to 10 (breaks a network's terms). The Team page shows, per collaborator, how many flags they got this week, this fortnight and this month, and for which rule. A review can name the exact content it judged: if the post was edited in between, the server refuses it and asks for a fresh look.