Skip to content

PRD Tool

PRD Tool is a CLI and AI assistant integration for creating, validating, scoring, and viewing Product Requirements Documents (PRDs) using a canonical JSON schema.

Features

  • Structured PRDs: JSON-based schema for consistent, machine-readable PRDs
  • Quality Scoring: Automated rubric-based scoring with actionable recommendations
  • Multiple Views: PM-focused and Executive summary views
  • AI Integration: Works with Claude Code (MCP) and Kiro IDE (Power)
  • Validation: Schema validation, ID format checking, and traceability verification

Quick Start

Install

go install github.com/agentplexus/agent-team-prd/cmd/prdtool@latest

Or build from source:

git clone https://github.com/agentplexus/agent-team-prd.git
cd agent-team-prd
go build -o bin/prdtool ./cmd/prdtool

Create a PRD

# Initialize a new PRD
prdtool init --title "User Authentication" --owner "Jane PM"

# Add content
prdtool add problem --statement "Users cannot securely access their accounts"
prdtool add persona --name "Developer Dan" --role "Backend Developer"
prdtool add goal --statement "Provide secure, seamless authentication"
prdtool add req --description "Support OAuth 2.0 login" --priority must

# Validate and score
prdtool validate
prdtool score

Use with AI Assistants

prdtool deploy --target claude

Then in Claude Code, the MCP server provides tools like prd_init, prd_add_problem, prd_score, etc.

prdtool deploy --target kiro-power --output ~/.kiro/powers/prdtool

The Power activates when you mention "PRD", "product requirements", or related keywords.

Documentation

Scoring Categories

PRDs are scored across 10 categories:

Category Weight Description
Problem Definition 20% Clear problem statement with evidence
Solution Fit 15% Solution addresses the problem
User Understanding 10% Personas and user journeys
Market Awareness 10% Competitive analysis
Scope Discipline 10% Clear goals and non-goals
Requirements Quality 10% Well-defined functional requirements
Metrics Quality 10% Measurable success criteria
UX Coverage 5% User experience considerations
Technical Feasibility 5% Technical constraints addressed
Risk Management 5% Risks identified with mitigations

Score Thresholds:

  • ≥8.0 → Approve (ready for implementation)
  • ≥6.5 → Revise (minor issues to address)
  • <6.5 → Human Review (significant gaps)
  • ≤3.0 → Blocker (critical issues)