What kind of moderation programs were people using then? Word filters? I can't imagine anything that isn't both easy to circumvent and likely to catch innocent posts merely quoting a "bad" word.
It wasn't like programmers were less creative or productive than they are now. The languages were fully capable of doing anything on a server you could do now. It was mostly C. Programmers were creative and smart, just like now, and could implement anything, albeit ground-up since there were fewer existing building blocks to reuse.
A lot of sophisticated anti-spam software depended on some sophisticated anti-anti-spam showing up first, and there is a lot more of that now for sure.
The type and quality of moderation is up to the moderation team. Might be automatic, might involve a review by a human for each post.
As someone else already mentioned, there are groups which were configured to be moderated but have no moderators at all to approve posts. Either they disappeared or there never were any. To participate in these groups you have to set the "Approved:" header in your post yourself. Obviously that would also be the way to circumvent the moderation process for any other moderated group. The countermeasure is that the moderators send out a cancel control message which instructs all the newsservers receiving it or at least those who honor cancel messages, to delete this specific message.
I wasn't "tech awakened" at this time, so I simply didn't know which programs were used.
What I can say is that plenty of spammers would have one or two posts on Newsgroups, before the mods would catch them and ban their email address.
Depending on the Newsgroup, I'm sometimes "new users" would have all of their posts reviewed before their posts were forwarded to everyone else. But after being added to some kind of whitelist, they could say whatever they wanted (well, until a flamewar caused them to go over the line and get banned).
So similar tools as moderators use today on Reddit / IRC / whatever. Or at least, it felt like that to me. Maybe it differed in implementation, but the overall effect was the same.
I don't think a "kill file" was the same as "moderation".
A kill file is "User X is tired of seeing User Y, so X stops reading posts from Y". Everyone else can still read Y.
A moderator is "Moderator is tired of seeing User Y, so Y cannot post to Newsgroup anymore". A much stronger action.
---------
Its been a few decades since I was part of any Usenet though. So maybe I'm mis-remembering the lingo. Or are you saying that the moderators used kill-files that somehow applied to the whole newsgroup?
Because... the later would make sense. But I really don't know how things worked back then.
There's a third layer: the servers. Newsgroups relied on servers propagating messages to other servers as well as delivering them to end users, and they could have various types of filters.
The same sort of model exists in modern distributed systems like Mastodon/ActivityPub. I can block a user from another server or an entire server individually, or my server admin can block them from communicating with anyone on my server.
What's missing that I think will be required for anything that gets popular enough is a means of sharing blocklists automatically, preferably with some machine-readable details so that they're useful even when servers have different rules (e.g. I want to subscribe to bans from Foo if the ban is for hate speech, but not for porn).
The machinery for a moderated newsgroup works by the news server you post an article to having a list of moderator email addresses for moderated groups. Instead of posting your article the server emails it to the moderator. The mod will then either just discard it, or else post it to their own news server with an Approved header to say it's moderated (and these days some crypto signing stuff). Then the post gets propagated as usual.
The "approve/deny/edit" stuff is handled by moderation software, which in principle can do anything you like.
