Cut to a finishable version 1

Your pitch decides what counts as the main job. Parts that serve that job score higher than parts that just sound nice.

Big parts with schemas and long prompts cost more. A small part that does a lot of the job always wins.

Nothing is deleted. Anything cut goes on a later list you can add in one press once version 1 is alive.

How much of what matters survived the cut. This should be high even when the number of parts drops a lot.

How much building you avoided for now. High work saved with high value kept is exactly the shape you want.

Each parked part carries its reason in plain words, so you can overrule us if you disagree.