[RFE] We should consider a test agent and validator that can attempt to resolve issues that occur when running the unit tests

[shawn-hurley]

I see two issues to overcome:

1. Running unit tests sounds good in theory, but many times, a fix in one class will cause unit tests to fail in another. This means that just unit testing the file that has been changed will not necessarily be sufficient. Running the full suite can take some time and would not really be feasible to run after each task.
2. A lot of configurations are usually involved in running unit tests, and this presents a challenge in making the agent work generically.