“Is there a heuristic we might use to identify and flag questionable papers?”

Article note: Not useful for editors, but for readers, the heuristic is "How recently was it published." The more recent, the more likely it'll be horseshit. There is, of course, some legitimate work going on in academia, but an ever increasing fraction is min/maxing the research/research artifact ratio so effectively that there isn't actually any research in it. Because that is what matches the incentive structure, and the norms have been steadily shifting to permit that kind of fraud.