The hard parts to design about this, imo, are:

- How to reconcile this with tests that execute many times with varying input data. You’d need some way to express requirements with specific inputs or shared inputs.

- Passing state between test dependencies.

- When, if ever, it’s fine to share step results between tests. If tests B and C require A, can you run A just once? Not always, but you should be able to when it’s safe.

I don’t think I’ve ever used a test framework that gets these things right.