Most AI users don’t struggle because the model is “bad”; they struggle because the request is underspecified. Your challenge is to transform a vague request into a reusable, production-quality prompt that consistently produces useful results.
Build a prompt that takes {{goal}} and turns it into a clear, personalized response for {{audience}}. Your prompt must give the AI enough context using {{context}}, respect {{constraints}}, follow the requested {{tone}}, and return the answer in {{output_format}}. It should also explicitly include {{must_include}} and avoid {{avoid}}.
The difficult part: your prompt should work even when some user inputs are incomplete. Instead of inventing facts, the AI should identify what is missing, ask only the most important clarification questions when necessary, and otherwise make clearly labeled reasonable assumptions. Include a final self-check against {{success_criteria}} before producing the answer.
Your submission will be judged on clarity, reusability, variable design, instruction hierarchy, output consistency, and how well it reduces generic or irrelevant responses.
Can you create one prompt that turns a messy human request into a response that feels specific, structured, and genuinely useful?