Create new command for testing existing model behavior

[russellb]

Create a new command to the bot that takes all of the added questions in the given taxonomy PR and captures existing answers from the model. The results should be published similar to the generate output we already do.

I'm not sure if lab provides a shortcut to this. Adding one would make this a little easier on us. Otherwise, I assume we can feed input into lab chat to do it a bit more manually.

We could also talk to the OAI endpoint directly, but I think it's better to go through lab to make sure we're using the same default prompt.

[russellb]

First implementation using lab depends on instructlab/instructlab#765

[russellb]

First implementation using lab depends on instruct-lab/cli#765

This is merged.

[russellb]

First implementation using lab depends on instruct-lab/cli#765

This is merged.

However, it won't actually work for the bot because lab test currently only works on Mac. Here's the issue for adding it to Linux: instructlab/instructlab#659