The only hard rule I'll agree to is no direct threats of violence.
Shame storms would be most effectively fought with better education, not moderators. Don't make judgements based on hearsay. Internet moderators don't help when a shame storm grows larger than the internet.
Bit from the article:
The solution, then, is not to try to make shame storms well targeted, but to make it so they happen as infrequently as possible. Editors should refuse to run stories that have no value except humiliation, and readers should refuse to click on them. It is, after all, the moral equivalent of contributing your rock to a public stoning. We should all develop a robust sense of what is and is not any of our business. Shame can be useful—and even necessary—but it is toxic unless a relationship exists between two people first. A Twitter mob is no more a basis for salutary shaming than an actual mob is for reasoned discussion. That would be true even if the shaming’s relics were not preserved forever by Google, making any kind of rehabilitation impossible.