AI-native product management skills that move at the speed your team needs.
Traditional PM frameworks were built for a world without AI — multi-week discovery sprints, day-long planning offsites, months of analysis before a decision. Other projects apply those same traditional models to AI-native development and call it progress. This project takes a different approach: every skill is designed from the ground up for how product management actually works when AI is part of the team.
Strategy, discovery, documentation, financials, analytics, delivery — every discipline compressed from weeks into hours.
💡 The philosophy: You already know lean — eliminate waste, move fast, validate early. We just gave it an AI engine. Every skill in this plugin is designed for hours and days, not weeks and months. AI handles the first 80%. You handle the last 20% — the 20% that determines whether the product succeeds.
What this project is not: A rehash of traditional PM playbooks with an AI label. Only skills that are optimized for AI-native speed will be added here. If a framework doesn't fit the emerging AI landscape — if it assumes weeks of manual synthesis or months of waterfall planning — it doesn't belong in this project.
In Claude Code, add the marketplace and install the plugin:
/plugin marketplace add mrthames/lean-pm-skills
/plugin install lean-pm-skills@lean-pm-skills-marketplaceThe first command registers this repo as a single-plugin marketplace. The second installs lean-pm-skills from it. After install, the skills are available in any Claude Code session.
To update later:
/plugin marketplace update lean-pm-skills-marketplacecurl -o CLAUDE.md https://raw.githubusercontent.com/mrthames/lean-pm-skills/main/CLAUDE.mdgit clone https://github.com/mrthames/lean-pm-skills.git🖱️ Cursor IDE Setup
mkdir -p .cursor/rules
cp lean-pm-skills/.cursor/rules/product-manager-guidelines.mdc .cursor/rules/See CURSOR.md for details.
| Skill | What It Does | Time |
|---|---|---|
| Discovery Process | Compressed discovery cycle — frame, research, interview, map, decide | 5 days |
| Product Strategy & Vision | Craft a product vision, north star metric, and strategy-on-a-page | 2-4 hours |
| User Research Methods | Plan and execute interviews, usability tests, or surveys with synthesis | 1-3 days |
| Pricing & Packaging | Value-based pricing analysis, tier design, and revenue modeling | 1-2 days |
| Build vs. Buy | Evaluate build in-house vs. vendor vs. partner with TCO analysis | Half day |
| Experiment Design | Lean build-measure-learn loop — hypothesis, cheapest test, validated learning | 1-2 hours |
| Skill | What It Does | Time |
|---|---|---|
| Sprint & Release Planning | Break epics into stories, model capacity, sequence work, flag risks | 1 hour |
| OKR & Goal Setting | Set quarterly OKRs — objectives, key results, alignment, stress-test | 2-3 hours |
| Go-to-Market Planning | Build a GTM plan — messaging, channels, enablement, launch timeline | 1-2 days |
| Roadmap Planning | Translate strategy and OKRs into sequenced, audience-specific roadmaps | 2-4 hours |
| Feature Sunset & EOL | Plan and communicate a feature deprecation or product end-of-life | Half day |
| Skill | What It Does | Time |
|---|---|---|
| Stakeholder Alignment | Get stakeholders aligned on a decision with a decision-forcing document | 2-4 hours |
| Tech Debt Communication | Quantify and communicate tech debt as a business investment case | 2-4 hours |
| Competitive Response | Assess and respond to a competitor move within 24-48 hours | 2-4 hours |
| Skill | What It Does | Time |
|---|---|---|
| PRD Writing | Write a complete PRD from problem statement to builder-ready spec | 1-2 hours |
| Epic Definition | Define an epic with strategic context, feature breakdown, and milestones | 1 hour |
| Feature Specification | Detail a feature with requirements, edge cases, and integration points | 30-60 min |
| Story Writing | Write user stories with testable acceptance criteria and splitting guidance | 15-25 min |
| Story Breakdown | Decompose large epics or problems into independently shippable slices | 1 hour |
| Skill | What It Does | Time |
|---|---|---|
| Financial Analysis | ROI, IRR, NPV, payback period, and cost-benefit analysis for product investments | 1-1.5 hours |
| Business Case | Complete business case — strategic rationale, financial model, risk assessment, recommendation | 2-4 hours |
| Skill | What It Does | Time |
|---|---|---|
| Analytics & Insights | Deep dive into product data — investigate questions, surface insights, build narratives | 1-2 hours |
| Metrics Review | Regular metrics review — spot patterns, flag anomalies, build the data narrative | 30 minutes |
| Growth & Retention Diagnostics | Diagnose churn and retention problems with cohort analysis and root cause identification | Half day |
| Post-Mortem | Structured, blameless post-mortem after a failure or incident | 2-3 hours |
Real scenarios. Real PM problems. Here's how Lean PM Skills helps you move faster.
You've heard the signal from customers, sales, or support — but you're not sure it's real. Traditionally, this means a 3-8 week discovery process: scheduling interviews, synthesizing feedback, mapping the opportunity space, and building a recommendation.
With Lean PM Skills: The Discovery Process compresses this into a 5-day cycle. Claude helps you frame the problem on Day 1, synthesize existing research on Day 2, prepare and debrief interviews on Day 3, map the opportunity space on Day 4, and build a build/validate/kill recommendation on Day 5. You still talk to real users — AI just eliminates the weeks of synthesis work between conversations.
What you get: Problem framing document, research synthesis, interview guides, opportunity map, and a decision recommendation with supporting evidence.
Every PM has been here. Competing priorities, unclear decision criteria, and a meeting on the calendar where someone needs to say yes or no. The traditional approach — back-channel conversations, lengthy slide decks, consensus-building over weeks — doesn't fit when you need a decision now.
With Lean PM Skills: Stakeholder Alignment helps you build a stakeholder map, draft a decision-forcing document that makes the trade-offs explicit, pre-empt the most likely objections, and prepare a decision memo — all in a 2-4 hour focused session. You walk into the room with a clear recommendation and the analysis to back it up.
What you get: Stakeholder map with positions and concerns, decision-forcing document, objection responses, and a decision memo ready for sign-off.
The dashboard is red but the explanations are vague. Is it onboarding? Is it a specific segment? Is it a competitor? The traditional approach means weeks of pulling data, scheduling deep dives with analytics, and chasing down anecdotes.
With Lean PM Skills: Growth & Retention Diagnostics gives you a structured diagnostic framework in half a day. Claude helps you design the cohort analysis, identify which segments are driving the change, build a root cause hypothesis tree, and design interventions — while you provide the product context and customer knowledge that data alone can't surface.
What you get: Cohort analysis framework, segment-level diagnosis, root cause tree, and a prioritized retention intervention plan.
The feature is almost done. Marketing wants messaging. Sales wants enablement. Support wants docs. Leadership wants a timeline. And nobody has coordinated any of it.
With Lean PM Skills: Go-to-Market Planning helps you build the full GTM package in 1-2 days: messaging framework, channel strategy, sales enablement materials, support preparation, and a week-by-week launch timeline. Claude generates the structure and drafts; you add the market knowledge and relationship context.
What you get: Messaging framework, channel strategy, enablement materials, support documentation, and a phased launch timeline.
Your Slack is blowing up. Sales is asking for talking points. Your VP wants a response by end of day. The traditional approach — schedule a competitive analysis sprint, convene a cross-functional working group — takes weeks you don't have.
With Lean PM Skills: Competitive Response helps you assess the move, evaluate your response options, draft sales talking points, and prepare a leadership brief — all in 2-4 hours. Claude does the rapid analysis while you apply the strategic judgment about what this actually means for your product and customers.
What you get: Competitive impact assessment, response options with trade-offs, sales talking points, and an executive brief.
It's Sunday night. Tomorrow's sprint planning. The epic is clear in your head but it's one big block in the backlog with no stories, no acceptance criteria, and no dependency map.
With Lean PM Skills: Sprint & Release Planning takes you from epic to sprint-ready in about an hour. Claude helps you break the epic into user stories with acceptance criteria, model team capacity, map dependencies, and identify the scope trade-offs you'll need to discuss with the team.
What you get: Broken-down stories with acceptance criteria, capacity model, dependency map, and scope trade-off recommendations.
Pricing decisions have huge leverage — and most teams spend either way too long or way too little time on them. The traditional approach involves weeks of competitive analysis, value metric identification, willingness-to-pay research, and financial modeling.
With Lean PM Skills: Pricing & Packaging compresses this into 1-2 days. Claude helps you identify value metrics, build a competitive pricing matrix, design tier structures, and model revenue scenarios — while you bring the customer knowledge and strategic context about where your product sits in the market.
What you get: Value metric analysis, competitive pricing matrix, tier design options, and revenue scenario models.
The engineers are frustrated. The codebase is slowing them down. But "we need to refactor" doesn't get budget. Leadership wants to know the business impact.
With Lean PM Skills: Tech Debt Communication helps you translate technical debt into a business investment case in 2-4 hours. Claude helps you quantify the velocity impact, calculate the risk exposure, frame the opportunity cost of not investing, and build a recommendation that speaks leadership's language — revenue, risk, and speed.
What you get: Business-framed investment case with velocity impact, risk quantification, opportunity cost analysis, and a specific recommendation.
You have dashboards. You have data. But turning numbers into a narrative that drives action takes time — and it's easy to miss the signal in the noise.
With Lean PM Skills: Metrics Review gives you a structured 30-minute weekly review process. Claude helps you pull the key metrics, spot trends and anomalies, compare against targets, and draft audience-specific narratives — whether you're updating your team, your VP, or your board.
What you get: Structured metrics review with trends, anomalies, target comparison, and audience-ready narrative.
Something went wrong — missed targets, quality issues, user backlash, or a full-on incident. The team needs to learn from it, but post-mortems often devolve into blame games or produce action items nobody follows up on.
With Lean PM Skills: Post-Mortem structures a blameless post-mortem in 2-3 hours. Claude builds the timeline from your notes, categorizes contributing factors objectively, extracts lessons, and drafts 3-5 specific, owned, time-bound action items. You facilitate the conversation and ensure the team actually learns.
What you get: Chronological timeline, contributing factors analysis, prioritized lessons, and 3-5 actionable follow-ups with owners and deadlines.
You have a hypothesis — maybe a new onboarding flow will improve activation, maybe a pricing change will increase upgrades. But building the full thing to find out is expensive. The traditional approach is either "just ship it and see" or a multi-week experiment design process.
With Lean PM Skills: Experiment Design walks you through the lean build-measure-learn loop in 1-2 hours. Claude helps you map your riskiest assumptions, design the cheapest valid test (fake door, concierge, A/B), set clear success/kill criteria, and monitor results. You define the hypothesis and make the ship/iterate/kill call when the data comes in.
What you get: Ranked assumption map, experiment brief with hypothesis and success criteria, monitoring framework, and a decision recommendation with validated learning.
You've got OKRs, a backlog, some commitments, and a strategy — but they're in different docs and nobody has a clear picture of what's shipping when. The traditional approach is a multi-day planning offsite or weeks of spreadsheet wrangling.
With Lean PM Skills: Roadmap Planning translates your strategy and OKRs into a sequenced roadmap in 2-4 hours. Claude synthesizes your inputs, proposes sequencing based on dependencies and value, and generates three audience-specific views — a detailed view for the team, a timeline view for leadership, and a theme-based view for stakeholders. You apply the sequencing judgment and political reality.
What you get: Sequenced roadmap with rationale, three audience-specific views (team, leadership, stakeholder), capacity analysis, and an explicit "not doing" list.
You know what to build and why, but it's all in your head or scattered across Slack threads and meeting notes. Engineering needs a formal spec, and you don't have time to spend days writing one.
With Lean PM Skills: PRD Writing takes you from problem statement to builder-ready PRD in 1-2 hours. Claude structures the problem framing, success metrics, scoped solution, user stories, edge cases, and readiness requirements. You provide the strategic context and make the scope calls. The output is structured so engineers and AI coding tools can start building directly from it.
What you get: Complete PRD with problem statement, success metrics, v1 scope, user stories with acceptance criteria, edge cases, dependencies, risks, and readiness checklist.
The initiative is approved but it's still one big block in the backlog. Engineering can't estimate it, design can't start exploring, and nobody's clear on what "done" looks like.
With Lean PM Skills: Epic Definition breaks it into a structured epic in about an hour — strategic context, feature decomposition, milestones, dependencies, and success metrics. Feature Specification then details each feature with requirements, edge cases, and integration points. Story Writing turns features into sprint-ready stories with testable acceptance criteria. The full hierarchy — epic to feature to story — in a single session.
What you get: Structured epic with features, milestones, and success metrics. Detailed feature specs with edge cases. Developer-ready stories with Given/When/Then acceptance criteria.
A metric moved — activation dipped, engagement spiked in one segment, conversion shifted after a release. The dashboard shows the what but not the why. You need to dig deeper before you can act.
With Lean PM Skills: Analytics & Insights structures a deep investigation in 1-2 hours. Claude helps you sharpen vague questions into specific investigations, design the analysis (segments, time ranges, comparisons), spot patterns in the data, and build an audience-specific narrative with recommendations. You bring the product context that turns numbers into meaning.
What you get: Sharpened investigation questions, structured analysis, pattern identification with root cause hypotheses, and an insights narrative with specific recommendations.
You know this initiative will pay off, but "trust me" doesn't get budget. Finance wants NPV, IRR, payback period, and sensitivity analysis. The traditional approach is weeks of back-and-forth with the finance team building spreadsheets.
With Lean PM Skills: Financial Analysis builds a rigorous financial model in about an hour — year-by-year cash flows, ROI, NPV, IRR, payback period, and three-scenario sensitivity analysis. When you need the full picture beyond just the numbers, Business Case wraps the financial model in strategic rationale, alternatives comparison, risk assessment, and an executive-ready recommendation in 2-4 hours.
What you get: Year-by-year cash flow model, five financial metrics (ROI, NPV, IRR, payback, cost-benefit ratio), sensitivity analysis across three scenarios, and an investment summary document ready for finance and leadership review.
Every skill follows the same model:
┌─────────────────────────────────────────────┐
│ 🤖 AI handles the first 80% │
│ Structure, research, drafts, frameworks │
├─────────────────────────────────────────────┤
│ 🧑 You handle the last 20% │
│ Judgment, context, relationships, decisions│
└─────────────────────────────────────────────┘
This applies across all three disciplines in AI-native product development:
- 🎯 Product managers use AI for artifacts and analysis, own the decisions
- 🛠️ Engineers use AI for implementation, own the architecture and quality
- 🎨 Designers use AI for exploration, own the user experience
No AI-generated artifact should leave your hands without your judgment applied.
| Document | What It Is |
|---|---|
| CLAUDE.md | Core behavioral guidelines — always loaded, governs how Claude operates as your PM partner |
| GETTING-STARTED.md | Your first session — from setup to producing a real artifact in 15 minutes |
| FRAMEWORKS.md | PM frameworks adapted for AI-native speed — JTBD, RICE, Kano, story splitting, SaaS metrics |
| TEMPLATES.md | Artifact templates — PRDs, epics, features, stories, experiments, stakeholder updates, readiness checklists |
| AGENTIC-WORKFLOWS.md | Autonomous agents on a cadence — competitive intel, analytics, feedback monitoring, and more |
| AI-NATIVE-WORKFLOW.md | How PM changes when your whole team uses AI — handoffs, readiness, enterprise tooling |
| EXAMPLES.md | Before/after demonstrations of each core principle |
- Start With the Problem, Not the Solution — Ground every artifact in the user problem and evidence
- Write for Builders — Specs should be unambiguous and directly actionable by engineers and designers
- Define Done Before Starting — Measurable success criteria before work begins
- Scope Ruthlessly — Ship the smallest thing that tests the hypothesis
- Stay Evidence-Based — Distinguish between what data says, what it suggests, and what is assumption
MIT