Why a contract
The long-form version of the thinking behind this tool: Why solo game devs never finish (and what to do about it). Read it if you want the reasoning before the tool.
"Scope creep" sounds like a process problem. It is really an emotional one. Mid-build, a new idea always feels more exciting than the half-finished thing in front of you. So you chase it. Then the next one. The project quietly becomes unfinishable, and one day you just stop opening it.
A written scope contract works because it moves the decision earlier, to a calm moment, before you are emotionally invested in feature number nine. When the shiny new idea arrives at hour 20, you do not argue with yourself. You check the contract. It already said no.
How to use it
- Fill it in before you write any code. The contract is a planning tool, not a postmortem.
- Watch the scope meter. If it is yellow or red, cut something now. Cutting on paper is free. Cutting after you have built it hurts.
- Keep a "version 2" list. Every idea you have mid-build that is not in the contract goes on that list instead of into the game. The list is not a graveyard, it is the sequel.
- Hand the contract to Claude. The build-order prompt gives you a step list scoped to exactly the contract, nothing extra.
- If you must change the contract, change it deliberately. Re-open this tool, adjust, re-read the meter. A contract you renegotiate on purpose is fine. One you ignore is not.