In addition to what I wrote in the other comment, a critical skill is to imagine the possibility that someone close to you may be manipulating you.

I am not saying that you must suspect all people all the time. But when strange things happen and you notice that you are confused, you should assign a nonzero value to this hypothesis. You should alieve that this is possible.

If I may use the fictional evidence here, the important thing for Rational!Harry is to realize that someone close to him may be Voldemort. Then it becomes a question of paying attention, good bookkeeping, gathering information, and perhaps making a clever experiment.

As long as Harry alieves that Voldemort is far away, he is likely to see all people around him as either NPCs or his party members. He doesn't expect strategic activity from the NPCs, and he believes that his party members share the same values even if they have a few wrong beliefs which make cooperation difficult. (For example, he is frustrated that Minerva doesn't trust him more, or that Dumbledore is okay with the idea of death, but he wouldn't expect either of them trying to hurt him. And the list of nice people includes also Quirrell, which is the most awesome of them all.) He alieves that he lives in a relatively safe bubble, that Voldemort is somewhere outside of the bubble, and that if Voldemort tried to enter the bubble, it would be an obviously extraordinary event that he would notice. (Note: This is no longer true in the recent chapters.)

Perhaps we should taboo "resistance to charisma" first. What specifically are we trying to resist?

Looking at an awesome person and thinking "this is an awesome person" is not harmful per se. Not even if the person uses some tricks to appear even more awesome than they are. Yeah, it would be nice to measure someone's awesomeness properly, but that's not the point. A sociopath may have some truly awesome traits, for example genuinely high intelligence.

So maybe the thing we are trying to resist is the halo effect. An awesome person tells me X, and I accept it as true because it would be emotionally painful to imagine that an awesome person would lie to me. The correct response is not to deny the awesomeness, but to realize that I still don't have any evidence for X other than one person saying it is so. And that awesomeness alone is not expertise.

But I think there is more to a sociopath than mere charisma. Specifically, the ability to lie and harm people without providing any nonverbal cues that would probably betray a neurotypical person trying to do the same thing. (I suspect this is what makes the typical heuristics fail.)

Would noticing when one is confused and writing the details down help?

Yes, I believe so. If you already have a suspicion that something is wrong, you should start writing a diary. And a very important part would be, for every information you have, write down who said that to you. Don't report your conclusions; report the raw data you have received. This will make it easier to see your notes later from a different angle, e.g. when you start suspecting someone you find perfectly credible today. Don't write "X", write "Joe said: X", even if you perfectly believe him at the moment. If Joe says "A" and Jane says "B", write "Joe said A. Jane said B" regardless of which one of them makes sense and which one doesn't. If Joe says that Jane said X, write "Joe said that Jane said X", not "Jane said X".

Also, don't edit the past. If you wrote "X" yesterday, but today Joe corrected you that he actually said "Y" yesterday but you have misunderstood it, don't erase the "X", but simply write today "Joe said he actually said Y yesterday". Even if you are certain that you really made a mistake yesterday. When Joe gives you a promise, write it down. When there is a perfectly acceptable explanation later why the promise couldn't be fulfilled, accept the explanation, but still record that for perfectly acceptable reasons the promise was not fulfilled. Too much misinformation is a red flag, even if there is always a perfect explanation for each case. (Either you are living in a very unlikely Everett branch, or your model is wrong.) Even if you accept an excuse, make a note of the fact that something had to be excused.

Generally, don't let the words blind you from facts. Words are also a kind of facts (facts about human speech), but don't mistake "X" for X.

I think gossip is generally a good thing, but only if you can follow these rules. When you learn about X, don't write "X", but write "my gossiping friend told me X". It would be even better to gossip with friends who follow similar rules; who can make a distinction between "I have personally seen X" and "a completely trustworthy person said X and I was totally convinced". But even when your friends don't use this rule, you can still use it when speaking with them.

The problem is that this kind of journaling has a cost. It takes time; you have to protect the journal (the information it contains could harm not only you but also other people mentioned there); and you have to keep things in memory until you get to the journal. Maybe you could have some small device with you all day long where you would enter new data; and at home you would transfer the daily data to your computer and erase the device.

But maybe I'm overcomplicating things and the real skill is the ability to think about anyone you know and ask yourself a question "what if everything this person ever said to me (and to others) was a lie; what if the only thing they care about is more power or success, and they are merely using me as a tool for this purpose?" and check whether the alternative model explains the observed data better. Especially with the people you love, admire, of depend on. This is probably useful not only against literally sociopaths, but other kinds of manipulators, too.

the challenge is having enough people in the group who are resistant to charisma.

Since we are talking about LW, let me point out that charisma in meatspace is much MUCH more effective than charisma on the 'net, especially in almost-purely-text forums.

Does CFAR or anyone else teach resistance to charisma?

Well, consider who started CFAR (and LW for that matter) and how he managed to accomplish most of what he has.

resistance to charisma?

Ex-cult members seem to have fairly general antibodies vs "charisma." Perhaps studying cults without being directly involved might help a little as well, it would be a shame if there was no substitute for a "school of hard knocks" that actual cult membership would be.

Incidentally, cults are a bit of a hobby of mine :).