Sorry beforehand for the long reply.

Initially, one of .ml's admins (who's also a Lemmy developer) manually excluded ani.social from the list of instances in the join-lemmy site, and defederated it from .ml. When requested to revert the change, he falsely claimed that the instance is "full of CSAM". Eventually, the other .ml admin + Lemmy dev reviewed the "evidence" brought by the first one, concluded "there's no CSAM" here, and reverted that change.

They kept ani.social defederated, but that's fine - .ml is strictly SFW, there's some NSFW content in ani.social, so it's consistent.

Some time goes by, and a user creates a thread about "Mahou Shoujo something" in the !anime .ml community. I don't like that series; but more importantly it is NSFW, so the discussion was removed by a third .ml admin (not a dev).

Then we (a few users, incl. me) started discussing the eventual migration of the comm to ani.social. Because we knew that issues like this would keep happening, it was the best for both sides. With those first and third admins finding low-hanging fruits to wreck the discussion across multiple threads, such as "it lists to a pedo instance" or "doxxing" people. Claims that are blatantly knowingly false, because:

• ani.social was linked in the sidebar of !anime@lemmy ml for ages, and the local admins never bothered with it. But "suddenly" it becomes an issue, concomitantly with people discussing the migration of a comm to another instance?
• one of the people discussing the migration brought the contradiction above to the admins' attention. And yet the link stayed there, even if the admins were in a position to change it. Showing that no, linking ani.social was not the real issue that prompted the removal of the discussion, but the discussion about emigrating from that instance.
• In no moment, the people talking about the admin actions referred to personally identifiable information, like "you're John Smith"; we solely associated the administrative actions with the usernames. And that was done in a neutral tone, with zero harassment from my knowledge. (Relevant tidbit: both admins clearly use pseudonyms.)
• To add injury, the third admin in question was grasping at straws to defend the necessity of an anime community in an instance about open source and privacy, in a way not too unlike spez' "I'm one of you! We snoos stand together!" babble.

From public PoV, the matter ends here: you have the .ml admin team enforcing hidden rules and taking users as cattle to be herded. From my PoV, it gets worse.

I used to moderate a large-ish comm there, called !snoocalypse, about Reddit's downfall. In that comm, users (including me, the mod) were consistently saying stuff like "Steve Huffman the greedy pigboy". And in no moment the .ml admins took action against it, or even contacted me to say "hey mod, don't let your users do that".

So, naming someone by their RL name to call him a "greedy pigboy" is not doxxing. But stating which admin took which action by their username, in a neutral way, is suddenly doxxing??? And there's no way that the admins never saw it, because they were often removing content there.

Of course, the content that they were removing was from another nature: posts criticising either the Russian Federation or the People's Republic of China, typically under the allegations that violated rules #1 and #2 (basically: bigotry and making people feel unwelcome, or something like this).

Don't get me wrong, my issue is not that they were removing that criticism. I probably wouldn't bat an eye if they had some written rule like "don't criticise the RF or the PRC here"; I do criticise both but I'd see it within their rights. My issue here is to distort what others users say to fit the rules being listed, in order to enforce some rule not being listed, that is literally Reddit admins tier behaviour.