I'm generally a little on the users-over-rules side, so it'll be hard for me to give an unbiased answer, but I'll present some research I did.
I put together a Data Explorer query to look for questions that were closed then reopened, filter only ones reopened under suspicious circumstances like reopens without edits, and print the list of close votes for each question.
The query can be found here:
https://data.stackexchange.com/aviation/query/edit/1037498 . There's a quick summary of count per user here too, but I'd advise you to look at the questions first before jumping to conclusions based on user statistics.
It's not very pretty but without some serious string magic that's the best I can get.
I realize there's some grey area and debate about these criteria for a "suspicious" or "overturned" close vote. We need some debate on both sides about the quality standards on this site, so I'll avoid calling these overturned closure votes "bad" or "incorrect".
These closures merit attention, but there aren't many of them
First, while every user is important, it's clear this isn't a huge problem in the scope of the site. I could only track down 32 cases of overturned question closures where the OP was a new user and the question was marked off-topic or unclear. Second, there don't seem to be any users whose close votes are usually overturned, as almost all voters have only a 3-10% rate of suspicious close-votes. The number of overturned close votes is only high (as in, greater than fifteen) for a handful of users who are voting to close and reopen hundreds of questions already.
Is the problem small frequency in large numbers, or large frequencies in small numbers?
We could focus on the power-users, because about half the overturned close-votes come from about a dozen users. However, with a few outliers, most power-users average about a 7% suspicious close vote rate- and that's including duplicates and other closure reasons not brought up by the original question. Most of these users are already voting to reopen hundreds of questions as well. Definitely room for improvement, but hardly a case of toxic users.
We could also focus on the ~20% of suspicious close votes that come from users who don't have many other closure votes. There are some definite outliers here, but I don't know any good way to address them. The low percentage coming from these infrequent users also suggest the already-high reputation threshold is sufficient.
Pareto graph of closure votes by user
My Two Cents
My advice for voters: in my experience obscure questions are actually less annoying and less likely to distract from better questions than questions on common topics. Vote based on whether an expert could answer the question, not just a generalist. Also, remember to keep an eye on questions you've voted to close and if the question is edited, remove your vote. Several of these questions were closed just minutes after being edited to fix the issues.
My advice for askers: I know you're likely not reading this, but if you are, please avoid using acronyms and explain things so that even the new guy on your job would understand.
Best way to address the situation
Unfortunately there are limited ways to address this. This thread is a great way to start discussion. However, most of the posts here come from users with low rates of suspicious closure like Peter Kämpf and ymb1. Mods could PM the outliers, but that seems to be standoffish and I know directed guidance on how to vote is normally reserved for extreme situations.
It'd be helpful if the language for "unclear what you're asking" made it clear that we're talking about clarity to a subject matter expert, not just to a generalist, but I can't think of a clearer syntax off the top of my head and I believe changing it would involve a change to all of Stack Exchange.
Here's question closures I cherry-picked where obscure subject matter is a factor