But if you show HN something... and it makes HN-reading KF users think "OH EXPLOITABLE!!!111!" so they post it to KF... they're basically going to test if your moderation works or not. If you only have some automated rules or even "AI" as the moderation, not humans, then they're very quickly going to work out what gets past your automated filter and what doesn't, faster than 4chan can make Microsoft's Tay say "Hitler did nothing wrong".

Then the KF users will gloat about the slur-ridden fish they've made.

Then KF-reading Sharty users will learn about it, the rest of the Sharty will join in the game.

Then they'll get angry that your "normie" userbase has given a Palestine flag fish named "River to the Sea" the most upvotes, and an Israel flag fish the most downvotes, they'll want to cheat and hack their own creations to the top. So they'll start looking for an admin interface.

When they find one, but find it needs credentials, the first place they'll go looking for password clues is all your socials, all your previous forums, basically your whole identity. And if they hit paydirt, of course they'll post whatever they find, because for them, doxxing people is fun too.