Twitter proof should check for pre-existing tweet

If for some reason the proof process fails in the middle, the twitter proving system should check the tweet stream to see if there's a pre-existing proof tweet, rather than requiring the user to tweet a new one. (I had to tweet five or six proofs before one went through.)

Thanks for a cool service!

No files, tests, or entry points are named. Start by locating the Twitter proving system and trace how it handles a proof process that fails partway through. Done means an existing proof tweet can be detected and used instead of requiring another tweet.

Written by the indexing model from the issue text.