I wish I had approximately double the hours in a given day, and also vastly more coding skill to help in meaningful ways.

It seems sort of odd that comments or messages reported for spam don’t offer any tools. Even a simple url pattern match that gives mods/admins the ability to click a checkbox to remember the link and take some predefined action in the future would be a rudimentary but effective option.

I mean, heck, it’s the fediverse. In my fantasy implementation of an anti-spam approach, it would be possible to federate these lists of untrusted links and assign consensus-based confidence scores for links generated from moderator actions across instances. (With options for instance admins to tailor their own trust scores of other instances, so that each instance can choose for themselves who they trust, just in case a couple rogue instance admins try to poison the spam filter.)
Same concept can be applied to banned accounts, although in that circumstance, I’d suggest they find a way to mask the email address when sharing it. Not that folks won’t just spin up a new email. But, you know. Something is better than nothing.

Hopefully that makes sense. I’m losing my mind with sleep deprivation.