How does the PRD differ from an Epic in this methodology?

A PRD provides multi-feature initiative context and high-level outcomes, while an Epic serves as a single-feature implementation contract with detailed acceptance criteria.

What is the primary output of the ls-research skill?

It produces a standardized Product Requirements Document (PRD) containing vision, personas, feature candidates, scope boundaries, and success metrics.

When should I use the ls-research skill?

Use it at the very beginning of a project when requirements are unclear, scope spans multiple features, or when you need to align stakeholders before writing technical specs.

Can I skip the research phase and go straight to coding?

Yes, if the feature is already well-defined and the scope is narrow, the Liminal Spec methodology allows you to skip directly to Phase 2 (Epic creation).

How does this skill handle AI context limits?

It follows the 'Context Economics' principle by isolating the research phase into a standalone artifact, allowing you to hand off a clean document to a fresh agent session for the next phase.

Product Research & PRD Creator

Name: Product Research & PRD Creator
Author: liminal-ai

byliminal-ai

•

Herramientas de Colaboración

Transforms vague product visions into structured Product Requirements Documents (PRDs) to guide AI-assisted development.

This skill implements Phase 1 of the Liminal Spec methodology, designed to bridge the gap between a raw idea and technical implementation. It helps developers and product owners explore product direction, align stakeholders, and define scope boundaries before committing to code. By following the 'Context Economics' philosophy, ls-research generates a self-contained PRD artifact that serves as a clean handoff for subsequent development phases, preventing 'context rot' and ensuring that AI agents remain focused and precise during implementation.

Características Principales

013 GitHub stars

02Automated PRD generation from vision or problem statements

03Feature candidate prioritization using Must/Should/Nice-to-have logic

04Context Economics optimization for clean AI agent handoffs

05Persona-based user workflow and pain point identification

06Explicit scope boundary definition to manage project creep

Casos de Uso

01Breaking down a broad vision into multiple concrete, implementable feature candidates

02Aligning multiple stakeholders on feature sets before technical scoping

03Exploring new product initiatives when initial requirements are unclear

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add liminal-ai/liminal-spec ls-research

For use in Claude.ai and ChatGPT

Download Skill