Skip to content

slgoodrich/agents

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AI PM Copilot for Claude Code

Answer product management's hardest questions in minutes, not days

Expert product management guidance powered by frameworks from Teresa Torres, Marty Cagan, April Dunford, and more - without hiring a product manager.

Right feature. Right users. Right time. Right reasons.

Not a project management tool. Not a generic AI assistant. A specialized copilot that combines:

  • Speed of AI (answers in minutes)
  • Depth of product management frameworks (Teresa Torres, Marty Cagan, April Dunford)
  • Context-aware reasoning (remembers your product, personalizes outputs)

Built for solo developers and small teams who need enterprise product management capabilities without the complexity.

Why AI PM Copilot?

You're a solo founder building your first product. You can code, but product management decisions feel like guesswork:

  • "Should I build feature X or Y first?"
  • "How do I validate this idea without wasting weeks?"
  • "Which users should I target?"
  • "When is the right time to launch?"

Generic AI (ChatGPT, Claude): Fast answers, but no product management frameworks or strategic depth Enterprise PM tools (Jira, Aha!): Deep features, but slow, complex, and expensive Manual PM work (Notion, templates): Flexible, but requires expertise you don't have

AI PM Copilot: Expert product management frameworks + AI speed = rigorous answers delivered in minutes.

Built on 16 proven methodologies from product management thought leaders. Context-aware intelligence that remembers your product. Developer-native experience inside Claude Code.

Built on Proven Product Management Frameworks

AI PM Copilot isn't generic AI advice - it's built on expert frameworks from product management thought leaders:

Research & Discovery: Teresa Torres (Continuous Discovery), Erika Hall (Just Enough Research)
Strategy & Goals: Marty Cagan (Empowered), Sean Ellis (Product-Market Fit), April Dunford (Positioning)
Execution & Planning: Jeff Patton (Story Mapping), Basecamp (Shape Up), Google (HEART/Goals)
Prioritization: Intercom (RICE), Dan Olsen (Lean Product Playbook)

16 proven methodologies. 100+ templates and frameworks. Expert rigor at AI speed.

How It Works

Just talk to Claude Code like you would a consultant:

You: "I'm building a SaaS analytics tool. Should I prioritize dashboards or API access first?"

→ product-manager routes to feature-prioritizer
→ Reads your product context from setup
→ Returns scored recommendations with rationale

Result: Expert RICE scoring in 2 minutes with explanation of why each feature ranks where it does.

Try it yourself:

  1. Install: /plugin install ai-pm-copilot
  2. Setup (optional, 5-10 min): /ai-pm-copilot:pm-setup
  3. Ask: "Should I prioritize feature X or Y?"

What's Included

8 Expert Product Management Agents - Available 24/7 for research, strategy, execution, and launch (1 orchestrator + 7 specialists)

1 Utility Agent - Automated codebase context discovery (context-scanner)

3 Multi-Agent Team Presets - Validation sprints, PRD stress tests, competitive war rooms

16 Proven Frameworks - RICE, JTBD, Lean Startup, Story Mapping, and more

100+ Templates & Assets - PRD templates, roadmap frameworks, interview guides

Context-Aware Intelligence - One-time setup, personalized outputs across all agents

Built on frameworks from Teresa Torres, Marty Cagan, April Dunford, Sean Ellis, Jeff Patton, and more.

Get Started (5 Minutes to Your First Expert Answer)

1. Install the Plugin

/plugin marketplace add https://github.com/slgoodrich/agents
/plugin install ai-pm-copilot

2. (Optional) Set Up Product Context

Spend 5-10 minutes setting up context once, save 80% of questions across all future interactions:

/ai-pm-copilot:pm-setup

3. Ask Your First Product Management Question

"Should I prioritize feature X or Y?"
"Design a user interview study for churn"
"Score these 5 features with RICE"

Get expert product management guidance in minutes, not days. No signup. No API keys. Just install and ask.

Use Cases

Daily Product Management Work

Feature validation:

"I want to add real-time collaboration. Is this worth building?"
→ Coordinates research-ops for validation, feature-prioritizer for scoring

Planning:

"Plan next quarter focused on retention"

Documentation:

"Write a PRD for SSO integration"

Strategic Planning

Quarterly Goals:

"Create Q2 goals for a pre-PMF SaaS product"

Market analysis:

"Research the competitive landscape for project management tools"

Roadmap:

"Build a now-next-later roadmap for next 6 months"

Research & Discovery

User research:

"Design a study to understand user onboarding pain points"

Validation:

"Help me validate this feature hypothesis with 10 user interviews"

Synthesis:

"Analyze these interview transcripts and identify top 3 insights"

Launch Planning

Go-to-market:

"Create a beta launch plan for our API product"

Positioning:

"Develop positioning for a new pricing tier"

Agent Teams (Experimental)

Multi-agent team presets for high-stakes product decisions. Multiple agents work in parallel, challenge each other's conclusions, and synthesize competing perspectives.

Requires Claude Code's Agent Teams feature. Check https://docs.anthropic.com/en/docs/claude-code for setup instructions.

Validation Sprint

/agent-teams:validation-sprint "AI-powered code review tool for solo developers"

Three agents investigate in parallel (idea-researcher, market-researcher, idea-skeptic), cross-examine each other's findings, and deliver a BUILD / DON'T BUILD / NEEDS MORE EVIDENCE verdict.

PRD Stress Test

/agent-teams:prd-stress-test path/to/prd.md

Three reviewers (market-fit, feasibility, scope) score your PRD independently, flag conflicts between their reviews, and deliver a READY TO BUILD / NEEDS REVISION / MAJOR REWORK verdict.

Competitive War Room

/agent-teams:competitive-war-room "Notion, Coda, Slite"

One researcher per competitor runs a deep-dive in parallel. Results are synthesized into a positioning map, battle cards, and strategic recommendations.

Agents

The toolkit provides 1 orchestration agent, 7 specialist agents, and 1 utility agent that work together to cover the complete product lifecycle:

1. product-manager (Main Router)

Your primary entry point - Intelligent agent that routes requests to specialized experts

Use for:

  • General product management questions and guidance
  • Multi-faceted problems requiring multiple perspectives
  • When unsure which specialist to consult

Example:

"I need to validate a feature idea, prioritize it, and plan delivery"
→ Routes to research-ops, feature-prioritizer, and requirements-engineer

2. market-analyst

Market research, competitive analysis, and positioning

Use for:

  • Competitive landscape analysis
  • Market sizing and opportunity assessment
  • Product positioning and differentiation
  • TAM/SAM/SOM calculations
  • Win/loss analysis

Example:

"Analyze our position against competitors in the project management space"
"Research competitive landscape for project management tools"

3. research-ops

User research, interviews, and validation

Use for:

  • User interview planning and synthesis
  • Usability testing
  • Problem/solution validation
  • Customer feedback analysis
  • Research program design

Example:

"Design a study to understand why 30% of users churn after first month"
"Synthesize these interview transcripts into themes and insights"

4. product-strategist

Vision, strategy, and goals

Use for:

  • Product vision and strategy
  • Quarterly and annual goals
  • Strategic roadmaps
  • North Star metric definition
  • Strategic trade-offs

Example:

"Create Q2 goals focused on reaching product-market fit"

5. roadmap-builder

Roadmap planning and sequencing

Use for:

  • Now-Next-Later roadmaps
  • Theme-based roadmaps
  • Quarterly planning
  • Feature sequencing and phasing
  • Milestone planning

Example:

"Build a 6-month roadmap for our mobile app launch"

6. feature-prioritizer

RICE, ICE, and prioritization frameworks

Use for:

  • RICE/ICE scoring
  • Kano model analysis
  • Value vs Effort matrices
  • Trade-off decisions
  • Backlog ranking

Example:

"Score these 10 features using RICE framework"

7. requirements-engineer

PRDs, specs, and user stories

Use for:

  • Product requirements documents (PRDs)
  • Technical specifications
  • User story writing
  • Acceptance criteria
  • Feature documentation

Example:

"Write a comprehensive PRD for dark mode feature"

8. launch-planner

Go-to-market and launch execution

Use for:

  • Launch planning and execution
  • GTM strategy and distribution channel selection
  • Beta programs
  • Launch messaging and positioning
  • Community launches (Product Hunt, Hacker News, Reddit)

Example:

"Create a launch plan for our new API product"

9. context-scanner (Utility Agent)

Automated strategic context discovery from codebase

Use for:

  • Auto-discovering features, tech stack, and integrations during setup
  • Refreshing stale context when product changes
  • Getting accurate baseline for competitive analysis
  • Fast onboarding on existing projects

What it discovers:

  • Features (from routes, pages, API endpoints)
  • Tech stack (from package.json, requirements.txt, etc.)
  • Integrations (from 3rd party packages)
  • Project scale (files, LOC, complexity)

What it doesn't do:

  • Code quality or architecture analysis
  • Performance or security assessment
  • Engineering recommendations

Note: This is a utility agent (not a PM specialist) that uses Haiku model for efficient file operations. It's automatically invoked by pm-setup when you have an existing codebase, or can be invoked by PM agents when they need current product state.

Example (automatic invocation):

/ai-pm-copilot:pm-setup
→ Detects codebase
→ "Scan codebase to auto-discover context? (yes/no)"
→ If yes: context-scanner runs, presents findings for validation
→ Time saved: 10-15 minutes vs manual entry

Agent Routing Pattern

The product-manager agent acts as an intelligent router that:

  1. Analyzes your request to understand the product management discipline(s) needed
  2. Routes to the appropriate specialist agent(s)
  3. Coordinates multi-agent workflows when needed
  4. Synthesizes outputs from multiple specialists

Simple routing:

"Help me prioritize features"
→ product-manager routes to feature-prioritizer

Multi-agent workflow:

"Validate this feature idea and create a launch plan"
→ product-manager routes to research-ops (validation)
→ Then to launch-planner (GTM strategy)
→ Synthesizes complete plan

Direct to specialist: If your question clearly fits one domain, Claude Code may route directly to the specialist without going through product-manager first.

Skills

Skills are knowledge frameworks that agents load progressively when needed, organized by the product development journey:

Stage 1: Validation (Before writing code)

Is this idea worth my time? Who are my competitors?

  • competitive-analysis-templates - SWOT, Porter's Five Forces, competitive positioning
  • market-sizing-frameworks - TAM/SAM/SOM, market validation
  • product-positioning - April Dunford framework, differentiation strategy

Stage 2: Understanding (Who am I building for?)

Who are my users? What problems am I solving?

  • user-research-techniques - Interviews, surveys, usability testing, JTBD
  • interview-frameworks - Interview design, questioning, synthesis
  • synthesis-frameworks - Affinity mapping, insight generation
  • usability-frameworks - Usability testing, Nielsen principles
  • validation-frameworks - Problem/solution validation, MVP testing

Stage 3: Scoping (What's the MVP?)

What features do I build? How do I avoid scope creep?

  • prioritization-methods - RICE, ICE, Kano, Value vs Effort
  • product-market-fit - PMF indicators, Sean Ellis test, MVP scoping

Stage 4: Speccing (Make Claude Code build it well)

Clear requirements, user stories, acceptance criteria

  • specification-techniques - Requirements writing, clarity
  • prd-templates - PRD formats and best practices
  • user-story-templates - User story formats, INVEST criteria

Stage 5: Planning (What's my roadmap?)

Phase 1, 2, 3 and milestone planning

  • roadmap-frameworks - Now-Next-Later, outcome-based roadmaps

Stage 6: Launch (How do I get users?)

Where do I launch? What's my messaging?

  • go-to-market-playbooks - GTM strategies, distribution channels
  • launch-planning-frameworks - Launch tiers, timelines, execution

Setup & Utility

  • pm-setup - Interactive wizard for product context initialization
  • codebase-scanning - Automated feature and tech stack discovery from code

Setup Command

/ai-pm-copilot:pm-setup

One-time context setup that makes everything better.

Creates 8 core context files that all agents reference:

.claude/product-context/
  ├── product-info.md          # Product description, users, value prop
  ├── business-metrics.md      # ARR, users, growth, retention
  ├── strategic-goals.md       # Goals, priorities, themes
  ├── current-roadmap.md       # Now-Next-Later roadmap
  ├── tech-stack.md           # Architecture, tech, constraints
  ├── customer-segments.md     # Personas, segments, ICP
  ├── team-info.md            # Team size, roles, constraints
  └── competitive-landscape.md # Competitors, positioning

Benefits:

  • Agents ask 80% fewer questions
  • Outputs are product-specific, not generic
  • Context reused across all agents and skills

Usage:

# First-time setup
/ai-pm-copilot:pm-setup

# Update context later
/ai-pm-copilot:pm-setup --update

Architecture

Agent-First Design

  • Two plugins - core PM toolkit (ai-pm-copilot) + multi-agent team presets (agent-teams)
  • Main orchestration agent + 7 specialist agents + 6 team agents with clear responsibilities
  • Intelligent routing through main product-manager agent
  • Progressive skills that load on-demand
  • Solo developer focus - no enterprise complexity

Agent-First Benefits

  • Natural conversation instead of memorizing commands
  • Agents collaborate automatically
  • Context flows between specialists
  • Simpler mental model
  • Install what you need: core plugin or full suite

Model Strategy

  • Expert-level strategic reasoning
  • Nuanced decision-making
  • Context-aware responses
  • High-quality outputs

Creating PRDs

Use requirements-engineer to create Product Requirements Documents with automatic template selection:

"Write a PRD for [your feature]"

Intelligent Template Selection: requirements-engineer uses complexity assessment to automatically select the right template:

How it works:

  1. Assesses 3 dimensions: Technical Complexity, Risk/Impact, Scope Breadth (0-10 each)
  2. Reads repository context (.claude/product-context/tech-stack.md, etc.) for informed scoring
  3. Applies decision thresholds based on average complexity score
  4. Communicates reasoning transparently with complexity breakdown

Templates:

  • Lean PRD (1-2 pages): Low complexity (avg < 4), < 1 week, simple features
  • Comprehensive PRD (3-5 pages): Moderate complexity (avg 4-7), standard features [DEFAULT]
  • Amazon PR/FAQ (3-6 pages): New products (detected by keywords)
  • Google PRD (5-10 pages): High risk (security/payments) or scale-critical (high complexity)

Example reasoning: "OAuth authentication" → Google PRD (Risk: 9/10 security-critical, triggers high-detail template)

You can always request a different template if the automatic selection doesn't fit your needs.

Storage: PRDs are automatically saved to prds/[feature-name].md using kebab-case naming:

  • prds/dark-mode.md
  • prds/csv-export.md
  • prds/user-notifications.md

Target Audience

Built for solo developers and small teams who need:

  • Enterprise product management capabilities without complexity
  • Conversational interface, not command memorization
  • Intelligent routing to the right expertise
  • Progressive learning and guidance

Perfect for:

  • Solo founders building their first product
  • Small teams without a dedicated product manager
  • Developers transitioning to product roles
  • Side projects and indie hackers

Not designed for:

  • Enterprise organizations with product management teams
  • Complex stakeholder management scenarios
  • Large-scale portfolio management
  • Multiple product lines with dependencies

Documentation

  • Architecture - Agent-first design and system overview
  • Agents - Complete agent reference and capabilities
  • Skills - Framework knowledge and progressive disclosure
  • Usage - Patterns, workflows, and best practices
  • Plugins - Plugin structure and organization

Inspired By

This plugin was inspired by wshobson/agents. If, like me, you're not a software developer by trade, I highly recommend checking out that repository to get better results from Claude Code.

Contributing

Contributions welcome! This toolkit grows with the community.

Ways to contribute:

  • Report bugs or issues
  • Suggest new features or improvements
  • Share usage examples and workflows
  • Improve documentation
  • Add new skills or assets

See CONTRIBUTING.md for detailed guidelines.

Requirements

  • Claude Code
  • Basic familiarity with product management concepts

Support

License

This project is licensed under the PolyForm Noncommercial License 1.0.0. See LICENSE for details.


Empowering solo developers and small teams with enterprise product management capabilities through conversational AI agents.

About

AI PM Workflows for Claude Code

Resources

License

Code of conduct

Contributing

Stars

Watchers

Forks

Contributors