prd_skill

Overview

This skill generates high-quality Product Requirements Documents (PRDs) for software and AI features, including executive summaries, user stories, technical specifications, and risk analysis. It enforces measurable requirements and a strict schema so teams get a single, actionable source of truth. Use it to translate product vision into development-ready specs.

How this skill works

I first run a discovery interview to fill gaps: core problem, success metrics, constraints, users, and non-goals. Then I synthesize inputs into a structured PRD following a strict schema: Executive Summary, UX & Functionality, AI requirements (if applicable), Technical Specs, and Risks & Roadmap. Outputs use concrete, testable acceptance criteria and a phased rollout plan.

When to use it

• Kicking off a new product or feature development cycle
• Converting a vague idea into engineering-ready requirements
• Documenting AI-powered capabilities with evaluation plans
• Aligning stakeholders on scope, metrics, and non-goals
• Preparing handoff materials for design, engineering, and QA

Best practices

• Always perform a discovery interview and ask at least two clarifying questions before drafting
• Write measurable acceptance criteria (times, error rates, thresholds) not subjective adjectives
• Define non-goals to protect scope and timelines
• Specify testing and evaluation plans for AI outputs, including datasets and benchmarks
• Iterate: deliver a draft and request targeted feedback on ambiguous sections

Example use cases

• Create a PRD for an intelligent search feature with citation accuracy and latency KPIs
• Document requirements for a recommendation engine including evaluation metrics and data privacy controls
• Translate executive request into a scoped MVP + roadmap for an AI-assisted editor
• Produce handoff-ready technical specs and integration points for frontend and backend teams
• Prepare compliance-focused PRD for a system handling PII with security and audit requirements