* Added `--repo` and `--name` flags for running only examples with a
specific name, or repo (substring matching)
* Fixed a race condition that caused hangs when running multiple
examples at the same repo and sha
* Fixed a bug where scoring was completely wrong because I had passed
the arguments to `apply_diff_to_string` in the wrong order
* The current evals now run quickly and without errors.

Release Notes:

- N/A