prd-to-build-plan

Restate the goal, set explicit non-goals, define minimal scope, then break the work into phases that each deliver something runnable. Add acceptance criteria per phase, surface dependencies and risks, and recommend a concrete first step.

Identify the core loop of the feature and make it the first phase, then order remaining capabilities behind it. Each milestone should be independently shippable and verifiable, such as data layer first, then create/edit flows, then polish.

A build plan should include a restated goal, explicit non-goals, in-scope and out-of-scope items, phased milestones, acceptance criteria per phase, known dependencies and risks, and a suggested first step to begin execution.

Yes. For vague requests it infers a reasonable scope, states assumptions explicitly, and proceeds rather than stalling. For huge PRDs it identifies the core loop and phases the rest behind it, and flags conflicting requirements with a proposed resolution.

Skip heavy planning for small features where a short plan suffices; the approach matches plan depth to the size of the work. The hard rule is to never start coding a large feature without a phased plan.