lean-startup
Design MVPs, validated learning experiments, and pivot-or-persevere decisions using Build-Measure-Learn
Install / Use
npx skills add wondelai/skills --skill lean-startupInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
Finance & AccountingSupported Platforms
Our assessment of lean-startup
lean-startup scores 84/100 on our quality scale, 69th of 102 Finance & Accounting skills we index.
Its SKILL.md is 17 KB long, well organised into 24 sections and no code examples: a thorough specification that gives an agent plenty to work with.
With 2,248 GitHub stars, it is one of the more widely adopted skills in the catalogue.
Maintenance, license and trust
- The repository was last updated about 2 months ago, so lean-startup is actively maintained.
- Our last check on 2026-11-05 found the source still online.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.
AI review by kimi-k2.7-code on 2026-10-01. Automated pattern scan on 2026-10-01. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
lean-startup compared with similar skills
All 4 of these similar skills score higher than lean-startup; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| lean-startup (this skill)by wondelai | 84 | 2.2k | 2mo ago | SKILL.md |
| taste-skillby Leonxlnx | 100 | 116.8k | 8d ago | CLAUDE.md |
| Claude-Code-Game-Studiosby Donchitos | 100 | 39.7k | 7d ago | CLAUDE.md |
| x402by browser-use | 100 | 121.0k | 3d ago | SKILL.md |
| algorithmic-artby anthropics | 100 | 177.9k | 55d ago | SKILL.md |
Frequently asked questions
- How do I install lean-startup?
- Run
npx skills add wondelai/skills --skill lean-startup. The install tabs above show the steps for each supported agent. - Which AI agents does lean-startup work with?
- It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is lean-startup safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is lean-startup still maintained?
- The repository was last updated about 2 months ago, so lean-startup is actively maintained.
Skill content
View source on GitHubname: lean-startup description: 'Design MVPs, validated learning experiments, and pivot-or-persevere decisions using Build-Measure-Learn. Use when the user mentions "MVP scope", "validated learning", "pivot or persevere", "vanity metrics", "test assumptions", "innovation accounting", "build-measure-learn", "minimum viable experiment", "should we pivot", "test a business idea cheaply", or "build the smallest version first". Also trigger when deciding what to include in a first version, measuring startup progress, or evaluating whether to change direction on a product bet. Covers innovation accounting and actionable metrics. For 5-day prototype testing, see design-sprint. For customer motivation analysis, see jobs-to-be-done.' license: MIT metadata: author: wondelai version: "1.4.0"
Lean Startup Methodology
A systematic approach to building startups and launching new products that shortens development cycles and rapidly discovers whether a business model is viable.
Core Principle
Entrepreneurship is a form of management. Success doesn't require a perfect plan or brilliant insight—it requires a systematic process for testing assumptions, learning from customers, and iterating rapidly. Most startups fail not because they couldn't build what they planned, but because they built the wrong thing: treat every plan as a set of hypotheses to falsify, and spend effort to eliminate waste and accelerate validated learning, not to execute a fixed roadmap.
Scoring
Goal: 10/10. Score a plan, experiment, or metric set by the five Quick Diagnostic rows—1 point each when the answer is yes, 2 points when it is also backed by evidence on the Validation Ladder (Level 3+):
- 9-10: every leap-of-faith assumption named and ranked by risk, the riskiest tested by a real MVP, actionable metrics defined, and explicit pivot criteria set before building.
- 5-6: a hypothesis and some MVP exist, but metrics are vanity or pivot criteria are undefined—decisions can't be made from the data.
- ≤3: waterfall thinking—building the full product first, asking customers what they want, or scaling before product/market fit.
State the current score and the lowest-scoring diagnostic row to fix next.
The Build-Measure-Learn Loop
The fundamental cycle: IDEAS → BUILD (product) → MEASURE (data) → LEARN (knowledge) → back to IDEAS.
Critical insight: Plan the loop backward:
- What do we want to learn? (hypothesis to test)
- How will we know if we learned it? (metrics)
- What's the minimum we can build? (MVP)
Goal: Minimize total time through the loop.
See references/build-measure-learn.md when planning an experiment—reverse-planning sequence, an experiment-design template, per-product-type loop examples, and the build/vanity-metric loop traps.
Validated Learning
Learning what customers really want through experiments on real behavior—not feature requests, surveys, or focus groups (people mispredict their own behavior). Measure what customers do, not what they say, and run experiments that could falsify your assumptions. Vanity wins (downloads, signups without engagement) are not learning.
The Validation Ladder:
| Level | Evidence | Strength | |-------|----------|----------| | 1 | "I think customers want this" | Weakest (opinion) | | 2 | "Customers said they want this" | Weak (stated preference) | | 3 | "Customers signed up for early access" | Medium (low commitment) | | 4 | "Customers paid a deposit" | Strong (real commitment) | | 5 | "Customers are actively using it" | Strongest (revealed preference) |
Target: Level 4-5 before building at scale.
Minimum Viable Product (MVP)
The version of a new product that allows maximum validated learning with the least effort. Not a prototype (technical feasibility), not a beta (quality), not a minimum marketable product—a learning vehicle, often embarrassingly small and low quality, and usually much smaller than you think.
MVP Types:
| Type | What It Is | When to Use | Example | |------|------------|-------------|---------| | Concierge | Manual service pretending to be automated | Test if solution is valuable | Food on the Table (manual meal planning) | | Wizard of Oz | Fake automation, manual backend | Test if automation is needed | Zappos (no inventory, bought shoes retail) | | Smoke test | Landing page + signup, no product | Test demand before building | Dropbox video (explained concept, measured signups) | | Single feature | One core feature only | Test which feature is most valuable | Twitter (just status updates) | | Piecemeal | Combine existing tools | Test workflow before custom build | Groupon (WordPress + email) |
Design questions: What's the riskiest assumption? What's the minimum that tests it? How do we measure whether it was validated?
See references/mvp-design.md when choosing and sizing an MVP—seven types in depth, a type-selection decision matrix, lower/upper sizing bounds, and the MVP Design Canvas.
Leap-of-Faith Assumptions
The assumptions that, if wrong, will cause your business to fail. Identify them, prioritize by risk (which failure would be fatal?), and test the riskiest first—never in order of ease.
| Assumption Type | Question | Test Method | |----------------|----------|-------------| | Value hypothesis | Do customers care about this problem? | Smoke test, concierge MVP | | Growth hypothesis | How will customers discover us? | Channel tests, referral experiments | | Retention hypothesis | Will customers come back? | Cohort analysis, engagement metrics | | Monetization hypothesis | Will customers pay? | Pre-orders, pricing tests |
Example—Dropbox: Leap of faith: "people will download and use a file sync tool." Test: explainer video before building scale infrastructure. Result: beta list grew from 5,000 to 75,000 overnight—demand validated.
See references/assumptions.md when mapping and ranking assumptions—the Impact-Uncertainty matrix, a prioritization scoring template, test methods per assumption type, and industry-specific assumption lists.
Innovation Accounting
Measuring progress when traditional metrics fail: revenue and customers start at zero, and vanity metrics look good without driving decisions.
1. Establish the Baseline
Measure current reality precisely, even if it's zero or embarrassing: conversion funnel (signup → active → retained → paying), engagement (DAU/MAU, session length, features used), economics (CAC, LTV, churn).
2. Tune the Engine
Run experiments to improve baseline metrics: A/B test pricing ($9 vs. $19/mo), onboarding completion rates, acquisition channels (SEO vs. paid vs. referral). Each experiment targets a measurable improvement through validated learning.
3. Pivot or Persevere
When tuning stalls, make the evidence-based call (criteria and pivot types below in Pivot or Persevere).
See references/innovation-accounting.md when building the baseline dashboard—funnel, cohort, and economics metric frameworks.
Actionable vs. Vanity Metrics
Vanity metrics make you feel good but don't change behavior; actionable metrics drive decisions and clarify cause and effect.
| Vanity | Why It's Bad | Actionable Alternative | |--------|-------------|------------------------| | Total signups | Always goes up, no context | % signup → active (conversion rate) | | Page views | Doesn't indicate value | Time on page, bounce rate | | Total users | Includes inactive/churned | Active users (DAU, WAU, MAU) | | Downloads | Doesn't mean usage | DAU/downloads (activation rate) | | Revenue | Without context | Revenue per cohort, LTV/CAC |
Three characteristics of actionable metrics: actionable (clear cause-and-effect, reproducible), accessible (simple, understood by everyone), auditable (underlying data can be checked).
Example: Vanity: "We have 100,000 users!" Actionable: "Channel X users retain 2x better than channel Y—double down on X."
Cohort analysis: Group users by signup date and track behavior over time—the only way to see whether the product is actually improving.
See references/metrics.md when building a cohort table or choosing what to track—a five-step cohort walkthrough and AARRR (Pirate Metrics) aligned with Lean Startup stages.
Pivot or Persevere
A pivot is a structured course correction designed to test a new hypothesis about the product, strategy, or engine of growth.
Pivot when: experiments repeatedly fail to validate hypotheses, metrics stay flat despite iterations, customer feedback contradicts the vision, or progress is too slow for the runway. Persevere when: metrics are improving (even slowly), clear learning is happening, and adjustments move the right direction.
Pivot Types:
| Pivot Type | What Changes | Example | |------------|-------------|---------| | Zoom-in | Single feature becomes the whole product | Instagram (photo filters from Burbn) | | Zoom-out | Product becomes a single feature | Flickr (photo-sharing from Game Neverending) | | Customer segment | Same problem, different customer | Groupon (activism platform → local deals) | | Customer need | Same customer, different problem | Potbelly (antique store → sandwiches) | | Platform | App ↔ Platform | YouTube (dating site → video platform) | | Business architecture | High margin/low volume ↔ low margin/high volume | Salesforce (software → SaaS) | | Value capture | Monetization model change | Android (paid → free + app revenue) | | Engine of growth | Viral, sticky, or paid model | Facebook (viral in colleges → paid advertising) | | Channel | How you reach customers | Salesforce (direct sales → self-service) | | Technology | Different technology, same solution | Apple (Intel → ARM chips) |
Cadence: Successful startups commonly pivot 1-5 times before product-market fit. Anti-pattern: "pivoting" without validating that the new direction solves the core problem.
See references/pivots.md when the data suggests a pivot—the data-driven pivot signals, a structured pivot-meeting agenda, leading indicators, and the Instagram/Slack/YouTube pivot stories.
The Three Engines of Growth
How a startup acquires and retains customers sustainably. Pick one engine, optimize it, then consider adding others—running multiple engines simultaneously dilutes focus and learning.
1. Sticky Engine of Growth
Retention-driven: growth rate = new customer acquisition rate − churn rate. Track churn rate, retention cohorts (30/60/90 days), and DAU/MAU. Fits SaaS, subscriptions, social networks. Strategy: improve the product until natural growth exceeds churn.
2. Viral Engine of Growth
Customers bring customers: viral coefficient = (% who invite) × (invites sent) × (% who join); above 1.0 means exponential, self-sustaining growth. Track the coefficient, viral cycle time, and referral attribution. Fits Dropbox, Hotmail, WhatsApp. Strategy: build virality into the product itself.
3. Paid Engine of Growth
Spend to acquire: requires LTV > CAC (target LTV/CAC > 3x). Track CAC, LTV, and payback period. Fits e-commerce and traditional businesses. Strategy: optimize until each customer's profit funds acquiring more.
See references/growth-engines.md when picking or tuning an engine—churn-reduction tactics, the K-factor and viral-loop design, LTV/CAC optimization, a channel-economics table, and the product-to-engine matching framework.
The Five Whys
Root cause analysis: when a problem occurs, ask "why?" five times, then invest proportionally at every level—not just the symptom.
Example—website went down:
- Why? Server ran out of
Truncated for display — read the full file on GitHub.
Related Skills
taste-skill
116.8kAgent skills that give coding agents taste: art direction, reference images and a browser check instead of generic AI pages
Claude-Code-Game-Studios
39.7kTurn Claude Code into a full game dev studio — 49 AI agents, 72 workflow skills, and a complete coordination system mirroring real studio hierarchy.
x402
121.0kSet up Browser Use Cloud payments with x402 — pay per request from a crypto wallet (USDC on Base mainnet), no signup or API key.
algorithmic-art
177.9kCreating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
