Build a clearly throwaway prototype to answer one design, logic, state, or UI question before committing to production code. Use when the user asks to prototype, mock up, try a design, sanity-check a model, or make something exploratory.
Instalación
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Build a clearly throwaway prototype to answer one design, logic, state, or UI question before committing to production code. Use when the user asks to prototype, mock up, try a design, sanity-check a model, or make something exploratory.
Prototype
Input
A design, state model, UI idea, interaction, algorithm, integration, or uncertainty to explore.
Use explicit input first; otherwise infer the question from context.
Safest default: define the question before coding.
Workflow
State the question. Define the single question the prototype must answer.
Choose the shape. Use a terminal, script, fixture, route, component, or mock UI that answers the question fastest.
Mark it throwaway. Name and place the prototype so it cannot be mistaken for production code.
Make it runnable. Provide one command or obvious path to exercise it.
Expose state. Print or render the relevant inputs, transitions, outputs, and edge cases.
Capture the answer. Record what was learned in the final report, docs, issue, or follow-up plan.
Delete or absorb. Remove the prototype or fold the validated idea into real code when the question is answered.
Output
Question answered
Prototype location and run command
Findings
What should be deleted, kept, or absorbed into production code
Guardrails
Do not add persistence unless persistence is the question being tested.
Do not polish beyond what is needed to learn.
Do not let exploratory code silently become architecture.
