If fedi had programmable kill lists, you could do something like deny <bad instance> in hosts or deny <bad instances> subsetOf hosts, where “<bad instance(s)>” is/are the instance(s) you don’t want posts from, “hosts” is the list of instance domains a given post is published from or that an author uses (maybe separate “postSources” from “authorHosts”), and “deny” is short for “deny if [a condition is met]”.