Flags on controversial topics
No comment from the flagger
No history of plagiarism, abuse, scams, spam, or flag wars for the flagged user
These seem wholly subjective and likely to lead to absolute mountains of false positives.
This doesn't seem viable on this criteria. Your bot would need to read minds for this to work.
I like the general idea, though.
No bot, this is all done by humans. There are plenty of mistakes that could happen, and a lot of gray areas to be explored, but hopefully the community can learn a lot from this experiment and be better for it.
Oh, sorry, I totally misread. In that case, consider my observations retracted! Looks like a good idea.