I've very quickly written "new" tests for the code I've added, and that all seems to work and hit the proper parts of the code. However.
Looking at the tests, I see that they're very superficial. There are very basic questions that I should be asking about the output of this code that I'm currently not sure how to ask. I remember considering some of them in the past, so I'd better take a quick look at the relevant module...
So, the Compendium, at least in this pre-future version where that's not actually what it's called, defines two diagnostic methods. These methods help to track down basic issues with the Compendium, and I should use them outside of the apparently just one test that they're currently in.
But I also need diagnostic stuff like: provide every command where the first argument ends with the given name. From there, I can check over the remaining arguments. This will let me confirm stuff like "commands will have everything installed required to actually run". Tomorrow or later, I'll try to work out what kind of interface I want on that. For now, I'm going to wind down again.