Àdánwò

Make ideas tangible so the user can discover, compare and shape the outcome they want. The result is a confirmed direction and clearer needs; the prototypes are disposable.

Frame enough to start

Reuse the known audience, intended outcome, constraints and user preferences. Identify what experiencing a prototype could clarify. Start with a provisional question when the need is still emerging; feedback may reshape it. Ask only for missing information that materially changes the first useful round.

Choose the smallest medium that makes the relevant experience truthful: a sketch for visual direction, an interactive flow for behavior, or a runnable API/CLI for ergonomics. Label simulations and never fake the dimension being explored. Reuse existing components and preview capabilities when useful.

Keep prototypes in temporary storage by default, or .qp/adanwo/ when persistence helps iteration. Stay within existing authority for edits and external effects; a prototype request does not authorize production changes, publication or adoption.

Explore distinct directions

Build the requested number of variants, or choose a useful range proportionate to the uncertainty and effort. Ten variants can be appropriate; ten cosmetic variations rarely clarify ten different possibilities. Give variants stable labels and explain the meaningful idea each explores. Keep them comparable on the current question and at enough fidelity for useful reactions.

A single prototype is sufficient when it can resolve the uncertainty. Use relevant design or implementation skills for craft while keeping this work disposable.

Match the experiment to its surroundings. For interface variants, use representative content, density and surrounding navigation in an isolated preview; an empty standalone screen can conceal the real design problem. Keep variant comparison easy without requiring a particular layout or saved URL state. Stub effects outside the question and avoid real mutations.