Malicious/toxic should be tolerated in DF.
I mean, we should be able to go balls to walls in there.
Post a disclaimer with an agreement that the users have to acknowledge before entering, even after their post count gets them there. They have to agree to the terms that if you can't handle it, don't go inside. Disclaimers work for South Park. They can surely work here. You're not liable for anything. Reports that are not for racism, threats or bigotry, result in the reporter getting a warning. Have one mod set aside just for DF. General Noise is a lot of ground to cover. Make a DF'er a mod of just the DF folder, to handle the few cases that are deemed rule breakers. Your problems would be solved then.
No bigotry or racism or threats. That doesn't need to fly anywhere.