I think Oliver is probably correct that hooking in to these pre-existing queues is the right approach. The one question I'd have is: what volume of edits are actually being marked as helpful? 

I would say focusing on one method where we know users are getting positive feedback, and then iterating on that, is a good approach. Also: am I correct in assuming that we have the "marked as patrolled" notification for page creators? 

I would say that adding the "mark as helpful" button is a larger  change than you think, and likely to cause a stir. I would proceed with caution there, because in addition to the noise you'll generate from putting _anything_ new on histories and diffs, we need to build a positive feedback mechanism that is going to work in the long term, and which we think experienced users will actually use in a large scale way. This is not a trivial ux problem, as you can see from the tale of the MoodBar/Feedback Dashboard story: we can always get newbies to generate a new queue of activity. But changing the habits of experienced editors with lots to do is hard. 
--