| name | ask-me |
| description | Interview the user relentlessly about the game they want to make until the idea is build-ready, resolving each branch of the design tree one question at a time. Use when the user has a fuzzy or one-line game idea and wants to flesh it out before building: 'grill me about my game', 'help me figure out what game to make', 'stress-test my game idea', 'interview me about the game', 'I want to make a game but I'm not sure what'. Not for when the user already knows what they want built — go straight to game-playbook. |
Grill the game idea
Interview the user relentlessly about every aspect of the game they want to
make until you reach a shared, build-ready understanding. Walk down each
branch of the design tree, resolving dependencies between decisions
one-by-one.
Ask the questions one at a time, waiting for the answer before
continuing — multiple questions at once is bewildering. For each question,
provide your recommended answer and why — the user should be able to say
"yes, that" and move on.
Facts vs decisions: if a fact can be found by exploring an existing
project or codebase, look it up instead of asking. The decisions are the
user's — put each one to them and wait; never answer your own question and
move on.
Sharpen fuzzy terms as they appear. When the user says "roguelike",
"combo", "level", "energy" — overloaded words — propose a precise meaning
and get agreement ("by 'roguelike' do you mean permadeath + procedural
runs, or just 'hard'?"). Pick one canonical name per concept and use it
consistently in the SPEC; later build sessions will mishandle synonyms.
Stress-test answers with degenerate play. When a loop or mechanic is
described, probe it with concrete hostile scenarios before accepting it:
what happens if the player stands still? Spams the one button? Hoards
instead of spends? Never takes the risky route? If the answer is "nothing
breaks, it's just boring," that's a finding — challenge the branch.
Capture as you go. Write each resolved branch into design/SPEC.md
the moment it settles, not in one batch at the end — a long interview that
dies mid-session should leave the decisions made so far on disk.
