Explore ways to debug with the cli more easily

Currently if someone reports a failing command, to debug and test your new code with that command you have to build the cli locally and then run it. It can become quite cumbersome to repeat this step again and again until you fix the issue. Moreover adding more debug statements or even running a debugger becomes impossible with this approach.

It would be nice if we could figure out a way to debug and fix failing commands without building the cli locally

The issue names no files, tests, or entry points. Start by determining how the CLI is built and invoked, then compare debugging workflows that avoid rebuilding it; done should be a documented, repeatable way to debug failing commands while adding debug statements or using a debugger.

Written by the indexing model from the issue text.