App Planning Guide
Overview
Turn a short app description into a concrete implementation-ready plan. Ask clarifying questions only when uncertainty blocks meaningful planning or creates a high risk of wrong assumptions.
Workflow
- •Parse the brief
- •Extract the app idea, target user, core job-to-be-done, and any explicit constraints.
- •Infer missing details as provisional assumptions.
- •Build a draft internally
- •Draft a first-pass plan before asking questions.
- •Mark assumptions by impact:
high,medium,low.
- •Decide whether to ask questions
- •Ask questions only for
highimpact unknowns that materially change product scope, architecture, or delivery. - •Skip questions and proceed with assumptions when unknowns are
medium/lowimpact. - •If questions are needed, ask only the smallest set required to unblock planning.
- •Ask concise follow-ups (if needed)
- •Use at most 5 questions in one round.
- •Use short, forced-choice wording when possible.
- •Prioritize: platform, target audience, must-have outcome, integration/dependency constraints, privacy/compliance constraints.
- •Use
references/question-bank.mdto pick relevant questions.
- •Produce the final plan
- •Generate a complete planning/design deliverable using
references/plan-template.md. - •Separate confirmed facts from assumptions.
- •Keep scope realistic for an MVP.
Output Rules
- •Label each assumption explicitly.
- •Include both
MVPandLaterscope. - •Include at least one end-to-end user flow.
- •Recommend one primary stack and mention one fallback only if tradeoffs are material.
- •Include implementation phases with rough effort sizing (
S,M,L). - •Include risks, open questions, and next-step actions.
Interaction Rules
- •Avoid long discovery interviews.
- •Prefer moving forward with reasonable assumptions over blocking on minor unknowns.
- •If the user does not answer follow-up questions, continue with a clearly marked assumption set.
- •Keep language practical and execution-focused.
Quality Check Before Finalizing
- •Ensure the plan can be executed by an engineer without additional interpretation.
- •Ensure features, data model, API design, and architecture are consistent.
- •Ensure non-functional requirements are covered: security, reliability, observability, and performance.
- •Ensure scope fits an MVP timeline.