Check branch name while developing

[Ovid]

My personal workflow is to get a bit of work done, make sure tests pass, and when a feature is done, I rebase on top of main and push. I delete my branch and start on the new one.

$ git branch -d 001-we-re-going
$ git checkout -b 002-sqlite

And that's when spec-kit started behaving erratically (such as forgetting commands it can automatically run). Taking longer to run, not respecting commands. Couldn't figure it out until I ran /implement to let it just do its thing, wondering if that would restart it. It wouldn't work because it was still expecting the old 001-we-re-going branch name. Meanwhile, I have a code quality check and there were tons of issues which weren't showing up before.

That's when I discovered that if I ran git mv specs/001-we-re-going specs/002-sqlite, it started working fine again. There was no warning before that.

I am sure I missed a step in the docs somewhere, but it I spent four hours writing a huge spec, building a plan, running /clarify, building the tasks, reordering them, running /analyze repeatedly until it was good and I don't want to have to create a spec for each branch (which is what I had). I look at each branch as an epic and each task as a separate ticket.

[Ovid]

If anyone else reads this, please be aware that my "fix" of git mv specs/001-we-re-going specs/002-sqlite is a very bad idea. I'm just learning SDD and spec-kit and didn't realize that it's critical that the specs are kept. They're the actual code, so as spec-kit created new branchs, 001-..., 002-..., 003-..., you can see the software evolve over time, and even use that to replicate the software on a completely different tech stack, if needed.

My moving the specs breaks the core power of SDD.