I see this as a client failure to implement relay configuration/outbox model.  Clients experimented and went down a path of free and open relays, with zero-conf relay configs.  They thought they could rely on WoT and mute lists, and put off implementing outbox.  With the outbox model, you can set what relay to use as your INBOX and OUTBOX.  The INBOX is the ONLY place you read comments and reactions from and you can setup a relay to do this like @Laeserin has already done.  She has a relay that limits posting to only her follows (using listr) + any other pubkeys she wants on there (setup by relay.tools).  The only additional thing relays can do is to add NIP42 auth so that you could have read restricted to that same set of pubkeys.  This would enable tiers of protection from anything nostr throws at you.

The benefit of outbox vs. mute lists, is the pre-emptive ability to control what you see.  Seeing stuff is disturbing, and muting it after seeing it, this is not good enough on it's own.  And unfortunately, even when someone sets up a relay config in one client that works, if they jump to a client that doesn't use relay configs properly they get a barrage of all that stuff they didn't want to see (like primal).

Anyway, I gotta run, but I wanted to help describe what I see on my end, since I spend a lot of time on the relay side of things and I am well familiar with comment bots (which are technically no different than humans following someone around and commenting on them).