I would wish to have some summary advice for the specific scenario of looking for a new postdoc position and trying to avoid ending up with a terrible supervisor. Let me narrow this question down by saying that the most important thing to avoid would be an unsupportive/pressurizing/controlling supervisor, while competence is much less important. As long as the supervisor is really nice and lenient ("laissez faire"), the rest does not matter. Then again, the real nightmare is of course when the supervisor is both extremely controlling and incompetent. (A controlling supervisor is not that bad if at least they do not pressure employees into crazy and pointless tasks/projects.) The field is experimental psychology (in case that matters).
So far I have just two ideas, but both are far from perfect.
(A) Check how they performed during their first 3-5 years in academia (i.e., from their PhD start), since high output (especially high-quality first-authored articles) during that time is I think a strong indication of competence. (Publications/achievements afterwards are always questionable as people for whatever reasons rising in ranks have increasingly more power and opportunities for potential swindling and extortion.) However, the problem with this is that it is only useful in case of high output, since low output does not necessarily mean incompetence (e.g., they may have had a terrible supervisor themselves), and it has in any case very little to do with being just a nice and decent person.
(B) Ask current and former employees. However, the problem with this is that these people, especially in case of a really terrible supervisor, may be extremely intimidated and/or even brainwashed. This is especially true for current employees, but, in case of former employees, they might simply be happy to have done with the place and might not want to get involved. Also, some may have become accomplices. In general, no one has a strong incentive for telling the truth (to a helpless stranger).