ClaudeKit PM - Product Management Framework

A comprehensive, AI-enhanced product management system implementing Long Chain-of-Thought methodology for systematic problem-solving, strategic planning, and product development.

🚀 Quick Start (5 minutes)

# Install the CLI
npm install -g pm-kit-cli

# Initialize in your project
cd your-project
pm-kit init

# Start using PM workflows
claude
/prd  # Create your first PRD

New to ClaudeKit PM? → Read the Getting Started Guide for a complete walkthrough.

One-Line Install:

curl -fsSL https://raw.githubusercontent.com/your-org/pm-kit-cli/main/install.sh | bash

📚 Documentation

• Getting Started - Complete setup guide with examples (15 min read)
• Installation Guide - Detailed installation instructions
• Commands Reference - All CLI commands explained
• Contributing - How to contribute to the project

Overview

ClaudeKit PM is a File System As Context framework that enables Product Managers to leverage AI for complex reasoning chains, systematic validation, and data-driven decision-making. It provides structured workflows, specialized agents, and comprehensive templates for all aspects of product management.

Key Features

• Long Chain-of-Thought Reasoning: Extended reasoning chains with 15+ systematic steps and validation checkpoints
• Multi-Agent Orchestration: Specialized agents for problem decomposition, PRD writing, research synthesis, prioritization, and analytics
• Systematic Validation: Multi-layer verification at each phase with backtracking capabilities
• Comprehensive Templates: Ready-to-use templates for PRDs, decisions, personas, and more
• File System As Context: Organized structure that enables efficient AI-enhanced workflows

System Architecture

claude-kit-pm/
├── CLAUDE.md                 # Entry point - guides Claude's behavior
├── .claude/
│   ├── workflows/            # Detailed workflow implementations
│   ├── agents/               # Specialized agent definitions
│   └── commands/             # Slash commands for quick invocation
├── prds/                     # Product Requirements Documents
│   ├── templates/
│   ├── active/
│   └── archive/
├── specs/                    # Technical specifications
│   ├── technical/
│   └── functional/
├── research/                 # User research and insights
│   ├── user-interviews/
│   ├── surveys/
│   ├── insights/
│   └── personas/
├── roadmaps/                 # Strategic roadmaps
│   ├── quarterly/
│   ├── annual/
│   └── strategic/
├── metrics/                  # KPIs and analytics
│   ├── dashboards/
│   ├── reports/
│   └── analyses/
├── decisions/                # Decision logs and rationale
│   ├── logs/
│   ├── templates/
│   └── reviews/
└── frameworks/               # Reusable PM frameworks
    ├── prioritization/
    ├── analysis/
    └── templates/

Core Workflows

1. Problem Decomposition

Command: /decompose [problem statement]

Systematically breaks down complex product problems using:

• Jobs-to-be-Done analysis
• Problem tree construction
• Root cause investigation
• Stakeholder impact mapping
• Dependency and risk analysis

Workflow: ./.claude/workflows/problem-decomposition.md Agent: ./.claude/agents/problem-decomposer.md

2. PRD Framework

Command: /prd [feature name]

Creates comprehensive Product Requirements Documents with:

• Multi-layer validation
• User story generation
• Acceptance criteria
• Technical requirements
• Implementation phasing
• Risk mitigation

Workflow: ./.claude/workflows/prd-framework.md Agent: ./.claude/agents/prd-writer.md

3. User Research Synthesis

Command: /synthesize [research data]

Transforms research data into actionable insights through:

• Thematic analysis
• Affinity mapping
• Sentiment analysis
• Pattern recognition
• Persona development
• Opportunity mapping

Workflow: ./.claude/workflows/user-research-synthesis.md Agent: ./.claude/agents/user-researcher.md

4. Feature Prioritization

Command: /prioritize [feature list]

Applies multiple prioritization frameworks:

• RICE scoring (Reach × Impact × Confidence / Effort)
• ICE framework (Impact × Confidence × Ease)
• Value vs. Effort matrix
• Kano model analysis
• MoSCoW method
• Weighted scoring

Workflow: ./.claude/workflows/feature-prioritization.md Agent: ./.claude/agents/prioritization-engine.md

5. Strategic Planning

Command: /strategy [initiative/goal]

Creates comprehensive strategic plans with:

• Vision and mission alignment
• OKR development
• Roadmap visualization
• Resource allocation
• Risk assessment
• Multi-stakeholder validation

Workflow: ./.claude/workflows/strategic-planning.md

6. Quick Decisions

Command: /decide [decision]

Facilitates structured decision-making:

• Decision framing
• Option generation
• Pros/cons analysis
• Risk assessment
• Decision documentation
• Validation planning

Workflow: Embedded in command

Specialized Agents

Problem Decomposer

Breaks down complex problems using first principles thinking and structured frameworks. Location: ./.claude/agents/problem-decomposer.md

PRD Writer

Creates comprehensive PRDs with multi-layer validation and systematic verification. Location: ./.claude/agents/prd-writer.md

User Researcher

Synthesizes research data into actionable insights and validates product decisions. Location: ./.claude/agents/user-researcher.md

Prioritization Engine

Applies multiple frameworks to feature backlogs with transparent rationale. Location: ./.claude/agents/prioritization-engine.md

Analytics Synthesizer

Transforms data into insights, tracks metrics, and provides predictive analytics. Location: ./.claude/agents/analytics-synthesizer.md

Comprehensive Workflows

• Problem Decomposition: ./.claude/workflows/problem-decomposition.md
• PRD Framework: ./.claude/workflows/prd-framework.md
• Strategic Planning: ./.claude/workflows/strategic-planning.md
• Feature Prioritization: ./.claude/workflows/feature-prioritization.md
• User Research Synthesis: ./.claude/workflows/user-research-synthesis.md
• Metrics & Analytics: ./.claude/workflows/metrics-analytics.md
• Stakeholder Management: ./.claude/workflows/stakeholder-management.md
• Cross-functional Coordination: ./.claude/workflows/cross-functional.md
• Risk Assessment: ./.claude/workflows/risk-assessment.md
• Documentation Standards: ./.claude/workflows/documentation-standards.md

Available Commands

| Command | Description | Agent | |---------|-------------|-------| | /decompose | Systematic problem decomposition | Problem Decomposer | | /prd | Generate comprehensive PRD | PRD Writer | | /synthesize | Synthesize research into insights | User Researcher | | /prioritize | Apply prioritization frameworks | Prioritization Engine | | /strategy | Create strategic plans | Multi-agent orchestration | | /decide | Quick decision framework | Decision facilitation |

Templates

PRD Template

Comprehensive template with all sections for product requirements documentation. Location: ./prds/templates/prd-template.md

Decision Log Template

Structured template for documenting decisions with rationale and validation. Location: ./decisions/templates/decision-log-template.md

Persona Template

Detailed template for creating evidence-based user personas. Location: ./research/personas/persona-template.md

Long Chain-of-Thought Methodology

Every workflow implements extended reasoning chains with:

1. Extended Reasoning: 15+ systematic steps with clear progression
2. Verification Checkpoints: Multi-layer validation at each major phase
3. Backtracking Mechanisms: Clear triggers for returning to previous phases
4. Error Correction: Specific protocols for detecting and fixing common mistakes
5. Multi-Layer Validation: Internal consistency, stakeholder alignment, strategic fit
6. Practical Templates: Ready-to-use formats and structures
7. Real Examples: Concrete illustrations of each concept
8. Measurement Systems: Clear metrics and success criteria

How to Use

Getting Started

1. Review CLAUDE.md: Understand the system's role and capabilities
2. Choose a Workflow: Select based on your current task
3. Use Commands: Invoke agents with slash commands
4. Follow the Process: Each workflow has detailed phase-by-phase guidance
5. Validate Outputs: Use built-in verification checkpoints
6. Iterate: Leverage backtracking for continuous refinement

Example Usage

Problem Decomposition:

/decompose Users are abandoning the checkout process

Create PRD:

/prd Shopping cart transparency feature

Prioritize Features:

/prioritize Q1 feature backlog

Strategic Planning:

/strategy AI-powered product recommendations

Make Decision:

/decide Should we build vs. buy payment processing?

Workflow Integration

The system is designed for seamless integration:

Problem Decomposition → PRD Framework → Prioritization → Strategic Planning
         ↓                    ↓               ↓                ↓
    Research Synthesis → Analytics → Cross-functional → Risk Assessment

Each workflow produces outputs that feed into others, creating a cohesive product management system.

Success Metrics

Track the effectiveness of ClaudeKit PM:

• Problem Solving: Root causes identified vs. symptoms treated
• PRD Quality: Stakeholder approval rate, rework reduction
• Prioritization: Predicted vs. actual impact accuracy
• Research: Insight actionability and product impact
• Strategic Planning: OKR achievement rate
• Decision Quality: Outcomes vs. predictions, regret rate

Best Practices

1. Start with Problem Decomposition

Before creating PRDs or prioritizing, ensure you understand the root problem.

2. Validate at Every Checkpoint

Don't skip verification steps - they prevent costly mistakes.

3. Document Decisions

Use decision logs to track rationale and enable learning.

4. Leverage Multi-Framework Analysis

Use multiple prioritization frameworks for robust decisions.

5. Iterate Based on Learnings

Use backtracking triggers to adjust when context changes.

6. Maintain Single Source of Truth

Keep documentation organized and updated in the file system.

7. Cross-Reference Continuously

Link related documents (PRDs, research, decisions, strategy).

Customization

ClaudeKit PM is designed to be customized:

• Adapt Workflows: Modify steps to match your process
• Adjust Weights: Change prioritization criteria weights
• Add Templates: Create organization-specific templates
• Extend Agents: Add specialized capabilities
• Modify Commands: Tailor slash commands to your needs

Error Correction Protocols

Every workflow includes:

• Common error patterns and detection methods
• Correction procedures for each error type
• Prevention strategies to avoid future occurrences
• Backtracking triggers for when to revisit earlier phases

Quality Assurance

Built-in quality checks at multiple levels:

Level 1: Completeness

• All required sections present
• No placeholder text
• All links working
• Data sources cited

Level 2: Clarity

• Unambiguous language
• Clear structure
• Defined terms
• Accessible to audience

Level 3: Accuracy

• Facts verified
• Assumptions validated
• Technical details reviewed
• Data quality confirmed

Level 4: Utility

• Serves intended purpose
• Enables decisions
• Referenced by team
• Achieves outcomes

Support and Feedback

For questions or improvements:

• Review workflow documentation in ./.claude/workflows/
• Check agent capabilities in ./.claude/agents/
• Consult templates in ./*/templates/
• Document learnings in decision logs

Continuous Improvement

ClaudeKit PM evolves through:

• Retrospectives on workflow effectiveness
• Tracking predictions vs. outcomes
• Refining frameworks based on results
• Building organizational knowledge base
• Sharing learnings across teams

License and Usage

This framework is designed for product management teams to enhance their workflows with AI-assisted reasoning and systematic validation.

Quick Reference

Most Common Workflows:

1. Problem Decomposition → PRD Framework → Cross-functional Coordination
2. User Research Synthesis → Feature Prioritization → Strategic Planning
3. Quick Decision → Risk Assessment → Documentation

When to Use What:

• New feature idea: Start with /decompose
• Validated problem: Use /prd
• Feature backlog: Apply /prioritize
• User feedback: Run /synthesize
• Strategic initiative: Create with /strategy
• Important decision: Document with /decide

Emergency Shortcuts:

• Quick decision needed: /decide
• Stakeholder alignment: Check workflows for communication templates
• Blocked by dependency: Review cross-functional coordination workflow
• Risk materialized: Consult risk assessment workflow

Version: 1.0 Created: 2024-11-21 Last Updated: 2024-11-21

For detailed implementation guides, refer to individual workflow and agent documentation.