I use beem. It works something like this

acc_abuse = Account(abuser)
flagged_comments = acc_abuse.history(start=start_time, stop=stop_time,only_ops=['vote'])

so I have known list of retaliate flaggers for which I collect all the votes between a certain time interval. Then I just have filter to get the ones with negative weight

for flagged_comment in flagged_comments:
            if flagged_comment['weight']<0:
                     print(flagged_comment['permlink'])

The above code prints the corresponding permlinks.

There are a lot more conditionals in the actual code and the regular owl code of course does all the upvoting itself :o)

The manual work of the bot is adding retaliate flaggers and white-listed users to the list. I don't think it is necessary to automise this part