That's why we write tests?

Before refactoring I often don’t know what depends on a function until tests fail or something breaks later.

I’m experimenting with a small local CLI that explains impact by walking the call graph.

Before going further I’m curious how do you personally reason about refactor safety?

Repo if you want to see what I’m testing: https://github.com/Anandb71/arbor