🧠 Agentic PM Skills

The PM operating system for AI-native product teams

9 Plugins · 93 Skills · 63 Commands · MCP on npm

Turn vague prompting into structured PM workflows for discovery, strategy, execution, AI product work, GTM, analytics, and PM learning.

Designed for Claude Code & Cowork · Works with MCP clients like Claude Desktop, Cursor, and Windsurf · Skills also work in Copilot, Gemini CLI, OpenCode, Codex CLI, and more

Getting Started · Starter Packs · Proof & Examples · Compatibility Matrix · FAQ

---

Agentic PM Skills packages proven PM frameworks into reusable skills and workflows so your AI helps you think more clearly, not just write faster.

Instead of prompting from scratch every time, you get structured commands for discovery, PRDs, prioritization, go-to-market, AI features, prototyping, and guided PM learning.

⚡ 30-Second Quickstart

1) Install once

MCP (recommended): add the server to your AI client.

claude mcp add pm-skills -- npx -y ai-pm-skills-mcp

For Claude Desktop, Cursor, and Windsurf setup snippets, see docs/getting-started.md.

2) Start with a real PM task

/discover Improve activation for first-time users
/write-prd Smart notification batching for enterprise teams
/ai-feature-launch AI copilot for support ticket triage
/plan-prototype Internal tool for customer interview synthesis

3) Keep moving through the workflow

Typical flow:

/discover → /strategy → /write-prd → /write-stories → /sprint → /plan-launch

---

🎯 Why teams use this

Generic AI often gives you long answers. This repo gives you structured PM thinking.

• Better decisions — use established PM frameworks instead of ad hoc prompting
• Faster output — go from idea to artifact quickly with pre-built workflows
• More consistency — get repeatable PRDs, research outputs, strategy docs, and launch plans
• Safer AI feature work — cover model selection, responsible AI, data strategy, and incidents
• Faster PM onboarding — use the guided learning plugin as a PM Foundations course

👥 Start here by role

If you are...	Start with	Then use
Product Manager	/discover	/strategy → /write-prd
PM-builder / prototype owner	/plan-prototype	/vibe-spec → /deploy-check
AI PM	/ai-feature-launch	/responsible-ai-review → /ai-metrics
PM leader / operator	/plan-okrs	/transform-roadmap → /write-update
New PM / learner	/course	/learn-foundations → /learn-capstone
Not sure where to start	/find-skill	Follow the recommended next command

🔟 Top starting commands

Command	Best for	What it gives you
/discover	New ideas	Discovery cycle: ideation → assumptions → prioritization → experiments
/lean-validation	Problem validation	Problem framing, hypotheses, assumptions, experiments, decision support
/write-prd	Shipping work	Full PRD from a feature idea or problem statement
/strategy	Product direction	Product Strategy Canvas
/plan-launch	GTM planning	Launch plan, channels, messaging, and metrics
/data-driven-discovery	Analytics-led PM work	Metrics → tracking → funnel → cohorts → experiments
/ai-feature-launch	AI features	End-to-end AI feature definition and launch planning
/plan-prototype	PM vibe coding	MVP scope, tool choice, build sequence
/course	PM learning	Guided PM Foundations Course
/find-skill	Routing	Best skill or command for the task at hand

📋 Copy/paste examples by persona

PM

/write-prd Enterprise SSO with SCIM provisioning for IT admins

PM-builder

/plan-prototype Customer interview repository with tagging, search, and AI summaries

AI PM

/ai-model-eval Which model should we use for support ticket summarization with strict PII controls?

Learner

/learn-foundations

---

🧰 Starter packs

Use these bundles when you want a simpler entry point than the full catalog.

Starter pack	Best for	Includes
PM Core	Most PMs	discovery + strategy + execution
AI PM	AI feature teams	AI PM + analytics + strategy
PM Builder	PMs prototyping with AI tools	vibe coding + execution + AI PM
PM Learning	New PMs and team onboarding	guided learning + execution + discovery

See full pack definitions in docs/starter-packs.md.

---

🏆 Proof, not just promises

This repo now includes buyer-style proof assets so people can evaluate it faster.

Mini case studies

• Discovery turnaround: onboarding drop-off → testable plan
• Execution example: vague feature ask → strong PRD
• AI PM example: risky AI idea → safer launch plan

Before/after comparisons

• Generic AI vs. PM Skills side-by-side examples

Example outputs

• 10 realistic output samples

Demo asset

• 90-second demo walkthrough script

Note: this repository includes proof-based examples and walkthrough assets. It does not invent customer testimonials or fake user quotes.

---

🏗️ How it works

graph LR
    A["You"] -->|/command| B["Workflow"]
    B --> C["Skill 1"]
    B --> D["Skill 2"]
    B --> E["Skill 3"]
    C --> F["Structured output"]
    D --> F
    E --> F
    F --> G["Suggested next step"]

Loading

Concept	What it means
Skill	A reusable PM framework or guided workflow
Command	A slash-command that chains one or more skills
Plugin	A domain pack such as discovery, strategy, execution, AI PM, or learning

---

📦 Available plugins

9 plugins · 93 skills · 63 commands covering the full PM lifecycle

Plugin	Scope	Skills	Commands	Best starting commands
pm-product-discovery	ideas, assumptions, interviews, experiments	12	6	/discover, /lean-validation, /triage-requests
pm-product-strategy	strategy, business models, pricing, decision quality	14	6	/strategy, /pricing, /challenge
pm-execution	PRDs, roadmaps, stories, sprints, communication	20	11	/write-prd, /write-stories, /sprint
pm-market-research	personas, segmentation, journeys, market sizing, competitors	6	3	/research-users, /competitive-analysis
pm-data-analytics	metrics, funnels, SQL, cohorts, A/B tests	7	7	/data-driven-discovery, /north-star, /design-funnel
pm-go-to-market	GTM strategy, ICP, growth, positioning, naming	7	4	/plan-launch, /growth-strategy, /marketing-plan
pm-ai-product-management	AI specs, model eval, responsible AI, prompts, incidents	8	6	/ai-feature-launch, /ai-model-eval, /responsible-ai-review
pm-vibe-coding	PM prototyping, tech decisions, code review, deployment	7	6	/plan-prototype, /vibe-spec, /deploy-check
pm-guided-learning	PM Foundations Course and guided practice	12	14	/course, /learn-foundations, /find-skill

For workflow-based browsing, see docs/workflows.md.

---

🛠️ Installation

1. MCP server — best default

ai-pm-skills-mcp exposes all 93 skills and 63 commands to MCP-compatible clients without cloning this repo.

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "pm-skills": {
      "command": "npx",
      "args": ["-y", "ai-pm-skills-mcp"]
    }
  }
}

Claude Code

claude mcp add pm-skills -- npx -y ai-pm-skills-mcp

Cursor / Windsurf

Use the same MCP server command in your MCP settings:

{
  "mcpServers": {
    "pm-skills": {
      "command": "npx",
      "args": ["-y", "ai-pm-skills-mcp"]
    }
  }
}

2. Claude marketplace install

Claude Cowork

1. Open Customize
2. Go to Browse plugins → Personal → +
3. Select Add marketplace from GitHub
4. Enter: tarunccet/pm-skills

Claude Code marketplace

claude plugin marketplace add tarunccet/pm-skills

Then install the plugins you want:

claude plugin install pm-product-discovery@pm-skills
claude plugin install pm-product-strategy@pm-skills
claude plugin install pm-execution@pm-skills
claude plugin install pm-data-analytics@pm-skills
claude plugin install pm-go-to-market@pm-skills
claude plugin install pm-market-research@pm-skills
claude plugin install pm-ai-product-management@pm-skills
claude plugin install pm-vibe-coding@pm-skills
claude plugin install pm-guided-learning@pm-skills

3. Skills-only install for other assistants

The SKILL.md files work in tools that read markdown-based skills.

Tool	Install path	Support
VS Code Copilot Chat	.github/skills/ or .github/copilot-instructions.md	Skills
Gemini CLI	.gemini/skills/	Skills
OpenCode	.opencode/skills/	Skills
Cursor	.cursor/skills/	Skills, or use MCP
Windsurf	.windsurf/skills/	Skills, or use MCP
Codex CLI	.codex/skills/	Skills
Kiro	.kiro/skills/	Skills

For a full matrix, see docs/compatibility-matrix.md.

---

🧭 Best starting workflows

Goal	Recommended path
Explore a new product idea	/discover → /strategy → /plan-launch
Validate a risky problem	/lean-validation → /setup-metrics → /write-prd
Turn a feature request into a build plan	/write-prd → /write-stories → /test-scenarios → /sprint
Launch an AI feature safely	/ai-feature-launch → /responsible-ai-review → /ai-metrics
Build a prototype with AI coding tools	/plan-prototype → /vibe-spec → /review-code → /deploy-check
Level up a PM team	/course → /learn-foundations → /learn-capstone

See more buyer-style use cases in docs/workflows.md.

---

✅ Release readiness and contributor trust

This repo includes:

• python3 validate_plugins.py for plugin validation
• TypeScript MCP server source in src/
• docs for onboarding, proof, FAQ, workflows, and compatibility
• a visible release checklist

---

❓ FAQ

Is this just a prompt library?

No. It is a structured library of PM skills and chained workflows with explicit frameworks, output shapes, and routing guidance.

Do I need to clone the repo?

No for MCP usage. Yes only if you want local plugin files, direct editing, or manual skills-only installs.

What is the fastest way to get value?

Install via MCP, run one of /discover, /write-prd, /ai-feature-launch, or /plan-prototype, then follow the suggested next command.

Which pack should I start with?

Most users should start with PM Core. If you are building AI features, choose AI PM. If you are building prototypes, choose PM Builder. If you are onboarding PMs, choose PM Learning.

Do commands work everywhere?

Commands work in Claude marketplace installs and through the MCP server. Skills-only installs work in more tools, but slash-command behavior depends on the host.

How do contributors validate changes?

Run python3 validate_plugins.py. For MCP/server changes, also run npm run build after installing dependencies.

See the full FAQ.

---

📖 About & attribution

Originally curated by Paweł Huryn (phuryn/pm-skills). Extended and maintained by Tarun Narang.

Framework inspiration and attribution remain preserved throughout the skills.

🤝 Contributing

Contributions welcome. See CONTRIBUTING.md.

📄 License

MIT — see LICENSE.

---

Made with ❤️ for the PM community

⬆ Back to top