Guides creation of comprehensive Product Requirement Documents (PRDs) for software projects through structured questioning and validation, then generates implementation task lists in JSON format. Use when users want to document a software idea, create specifications for development, plan a new application feature/bug, or break down requirements into actionable tasks. Transforms ideas into implementation-ready documents with verifiable pass criteria.
Detected risks:
The prd-creator skill turns a rough software idea into a comprehensive Product Requirement Document (PRD) and then into an implementation-ready task list. It is aimed at less-experienced developers who need help translating intent into clear specifications, using structured questioning to fill in architectural gaps before any code is written.
It runs a multi-part process: gather a description, then create a detailed PRD.md through clarifying questions covering app overview, target audience, success metrics, competitive analysis, core features and user flows, technical stack, prerequisites, security, and assumptions. It verifies prerequisites and seeds a .env.local with placeholder values only, produces an executive summary for approval, and then analyzes the approved PRD to generate a JSON task list. That list starts with a mandatory prerequisite-verification task, keeps each task small (about ten minutes), categorizes tasks by type, and gives each explicit pass criteria, alongside a concise project SUMMARY.md.
It is most useful when planning a new application, feature, or bug fix and wanting verifiable, developer-ready specifications. It deliberately does not generate code, and never writes real secrets, using placeholders and asking the user to fill real values manually.
A comprehensive PRD.md, a JSON implementation task list with verifiable pass criteria, and a concise project SUMMARY.md.
Primarily beginner-level developers who want to turn a software idea into clear, implementation-ready specifications.
No. It focuses on documentation and task specification and deliberately does not generate code.
As a JSON list starting with a mandatory prerequisite-verification task, with each task small enough to finish in about ten minutes, categorized by type, given explicit pass criteria, and initialized as not yet passing.
It only writes placeholder values to .env.local and never puts real secrets in the PRD, tasks, chat, or logs, asking the user to fill real values manually.
Quick Setup:
.claude/skills/Repository
pageai-pro/ralph-loop