Open Source · MIT License · 15 Claude Agents

The AI Operating System
Every Founder Needs.

Mark Zuckerberg is building a CEO AI agent for himself. Now every founder can. 15 Claude agents on the 3P3R Method® — a Human-Assisted Autonomous Organization (HAAO) where the agents do the work and you make the calls.

By Ludovic Bodin · Author of Atomic Scaling · Inspired by Garry Tan's GStack
Install via Claude Code (recommended)
/plugin marketplace add kalibrio/atomic-scaling-os
Never used Claude Code? The full walkthrough below takes 5 minutes, no terminal needed — or read it on GitHub →
15
Claude Agents
6
3P3R Pillars
MIT
License
HAAO
Org Model
30s
Install
Signal from the top
WSJ · 2026
"Mark Zuckerberg is building a CEO agent to help him do his job."
Meta is reducing management layers, maximizing individual contributors, and pushing AI-native operations — the same principles behind Atomic Scaling OS.
Read the WSJ article →
The HAAO shift
"The agents do the work. You make the calls."
If the world's most-resourced CEO is building his own AI operating system, the question for every other founder is simple: why aren't you? Atomic Scaling OS gives you the same leverage — free, open source, running in 30 seconds.
The pattern
Zuckerberg is building a CEO agent that retrieves answers directly, cutting communication layers
Meta is maximizing individual contributors, minimizing managers — the Atomic Scaling model
The competitive threat is AI-native startups moving faster with smaller teams

Most founders scale by hiring. That's the wrong lever.

Supercell built $3B with 300 people. Valve reached $10B+ with 400. ClickFunnels hit $200M revenue with 3 salespeople. Now Zuckerberg is building AI agents to cut his own management layers. The AI-native startups competing with you are doing the same. Here's what's holding you back.

✗ PEOPLE

Hiring Before Systemizing

You bring in people to solve problems a playbook could handle. Every hire adds coordination overhead that slows you down.

✗ PREDICTION

Decisions Without Data

You run on instinct instead of experiments. No hypothesis framework means no calibration — you don't know what's actually working.

✗ PLAYBOOK

No Finite Game

Your mission is fuzzy, your method is vague, and your 30-day priorities don't connect to your 5-year goal. Nothing compounds.

✗ REVENUE

Wrong Monetization Model

You're leaving money on the table between your free and paid tier. The Minnow/Dolphin/Whale gap is costing you LTV.

✗ REACH

Random Acts of Marketing

No systematic acquisition engine. CAC by channel is unmeasured. Growth is either feast or famine with no predictable pipeline.

✗ RETENTION

Retention as an Afterthought

Most companies that plateau do so because retention is broken, not acquisition. You're investing in the wrong end of the funnel.

Six pillars. Two agents each. One operating system.

Every scaling decision maps to one of six pillars — three for organizational capability, three for growth. Each pillar has two specialist agents running on scheduled loops.

"The companies that scale the fastest aren't the ones that hire the most. They're the ones that empower the fewest people to do the most."— Ludovic Bodin, Atomic Scaling
People
Stay small, scale output

Org architecture + culture + Dream-100 talent

OP = i × T scoring, magical pairs, EPIC evaluation, Supercell autonomous cells. 7-dimension culture pulse with weekly burnout detection. D100 (Talents) — the deliberate, slow-build pipeline of the hundred people who, if any one joined, would change the company's trajectory.

Prediction
Speed of iteration = speed of scaling

Experiments + superforecasting + premortems

Pendulum Framework for fast iteration, kill thresholds, Tetlock superforecasting, calibration scoring. Plus Klein-style premortems on high-stakes decisions.

Playbook
Create a finite game and play to win

Mission design + operational rhythm + postsuccess

MUM Model, Moonshot Thinking, Merlin Exercise (5yr→1yr→30day). Weekly scorecards and 12-quarter planning. Postsuccess analysis to reverse-engineer the 100x version of any strategy.

Revenue
Offer your basic service for free

Freemium design + unit economics

Ikigai of Freemium (7-step process), Minnows/Dolphins/Whales segmentation. LTV/CAC optimization — recoup in 6 months, 3x in 18.

Reach
Serve the largest total addressable market

Acquisition funnels + growth tracking

Dream 100 partnerships, OBE traffic model (Own/Buy/Earn). CAC by channel, 12-week growth sprints.

Retention
More important than reach or revenue

LOVE Machine + community engine

Live product dev, editorial calendar, segmented offers, re-engagement loops. Build community as a retention moat with <5% annual churn.

Fifteen specialists. All slash commands. All free.

This is a Human-Assisted Autonomous Organization (HAAO). The agents run your growth operating system on scheduled loops, surfacing only the decisions that need a human.

PeopleStay small, scale outputPillar 1
/people-team-architect
Org Designer

Design high-output teams

OP = i × T scoring, magical pairs, EPIC evaluation, Supercell autonomous cells, Valve 2.5x rule. Tells you who to hire — and who not to.

/people-culture-pulse
Culture Analyst

Monitor culture health

7-dimension pulse survey. Culture Health Score (0–100), burnout detection, trust metrics. Weekly alerts before problems become crises.

/people-d100-talents
Talent Pipeline Manager

Build the Dream-100

Paste any LinkedIn URL and get a paste-ready row for your D100 Talents sheet. 13-column schema, 5-step funnel: Followed → Connected → Exchanged → Offered → On-boarded. The deliberate pipeline of the hundred people who would change your trajectory.

PredictionSpeed of iteration = speed of scalingPillar 2
/prediction-hypothesis-engine
Experiment Lead

Run the H→M→C cycle

Pendulum Framework for fast iteration, kill thresholds, experiment management. Stop guessing — start measuring.

/prediction-forecast-analyst
Superforecaster

Probability-based planning

Tetlock superforecasting methods. Calibration scoring and scenario planning. Know what's coming before it arrives.

/prediction-premortem
Failure Investigator

Find the blind spots before launch

Gary Klein's premortem method. Assumes failure 6 months out, works backward to find every reason why, returns a revised plan. Run before any high-stakes commitment.

PlaybookCreate a finite game and play to winPillar 3
/playbook-mission-designer
Chief Strategist

Define your playbook

MUM Model (Mission/User/Method), Moonshot Thinking (10x > 10%), and the Merlin Exercise — work backwards from 5 years to what you do in the next 30 days.

/playbook-rhythm-keeper
Operations Lead

Maintain business cadence

Weekly scorecards, 12-quarter planning, Beast/Least prioritization. The rhythm that keeps scaling companies from losing focus.

/playbook-postsuccess
Success Architect

Reverse-engineer the 100x version

The strategic counterpart to /prediction-premortem and a deep-cousin of the Merlin Exercise. Assumes the initiative became exponentially more successful 12 months out, then surfaces the hidden leverage, compounding loops, and asymmetric advantages that made it inevitable. Returns a strategy with bigger ambition and stronger compounding.

RevenueOffer your basic service for freePillar 4
/revenue-freemium-architect
Monetization Designer

Design your freemium model

Ikigai of Freemium (7-step process). Minnows/Dolphins/Whales segmentation (R = M + D + W). Find where free meets paid.

/revenue-optimizer
Unit Economics Analyst

Optimize LTV/CAC

Pricing strategy, scale readiness scoring. Recoup CAC in 6 months, 3x in 18 months — or know exactly why you're not.

ReachServe the largest total addressable marketPillar 5
/reach-funnel-builder
Growth Architect

Build the Peak Sales Machine

Dream 100 partnerships, OBE traffic model (Own/Buy/Earn), follow-up funnels, anchoring. Systematic acquisition, not random acts of marketing.

/reach-growth-tracker
Acquisition Analyst

Track CAC by channel

12-week growth sprints, acquisition experiment management. Know exactly what's working and what to kill.

RetentionMore important than reach or revenuePillar 6
/retention-love-machine
Retention Engineer

The LOVE Machine

Live product development, Organized events (editorial calendar), Very special sales (segmented offers), Engagement/re-engagement loops. Build retention into your DNA.

/retention-community-engine
Community Architect

Community as a retention moat

CMX's 6 principles. Move users from "I use X" to "I AM an X member." Identity-driven retention targeting <5% annual churn.

Define → Measure → Iterate → Grow → Retain → Reflect

The agents run in the cadence a scaling company runs. Each feeds into the next. Nothing falls through the cracks — every agent knows what came before it.

FrequencyWhat happens
DailyExperiment scan, LTV/CAC monitoring
WeeklyCulture pulse, scorecards, metrics digest, growth tracking, retention check, iteration review
Bi-weeklyFunnel conversion audit, editorial calendar review
MonthlySegment analysis, calibration scoring, community health, revenue deep review
QuarterlyOrg design review, Merlin Exercise (5yr→1yr→30day), community health report
You: /playbook-mission-designer
Agent: Let's define your playbook. What's your company's mission in one sentence?

You: We help small restaurants get more local customers.
Agent: [runs MUM Model analysis]
Your mission is clear but your method is vague. Let me run the Merlin Exercise — where do you want to be in 5 years?
You: 10,000 restaurants across Europe.
Agent: MERLIN EXERCISE COMPLETE:
5yr: 10K restaurants, €50M ARR 1yr: 200 restaurants in France, PMF confirmed 90d: 20 paying restaurants, 2 cities 30d: Launch MVP, onboard first 5 restaurants in Lyon → Feeding into /revenue-freemium-architect for pricing model...

Running in 5 minutes. No terminal needed.

Never used Claude Code? Follow these exact steps. Anything weird, jump to the troubleshooting section in the full guide.

Step 1 — Download Claude Code Desktop
# In your browser, go to:
https://claude.com/code

# Click "Download for Mac" (or your OS).
# Open the .dmg, drag Claude Code into Applications.
# Open it. Sign in with your Claude account.
Step 2 — Open any folder as a project
# In Claude Code: File menu → Open Folder…
# Pick any folder — even an empty one on your Desktop.
# The skills write HTML reports into this folder.
Step 3 — Install the plugin (two commands)
# Type these into the chat box at the bottom
# of the Claude Code window — like ChatGPT.

/plugin marketplace add kalibrio/atomic-scaling-os

/plugin install atomic-scaling-os@kalibrio/atomic-scaling-os
Step 4 — Confirm 15 skills are loaded
# Type "/" in the chat box. A menu pops up.
# Start typing "play" — you should see:

/playbook-mission-designer
/playbook-postsuccess
/playbook-rhythm-keeper

# If you see those, you're done installing.
Step 5 — Run your first skill
/playbook-postsuccess

# I'm running a 6-week paid cohort for founders on
# operationalizing 3P3R with AI agents. $2k/seat,
# 30 founders. Target: $500k–$5M ARR founders.

# Press Enter. Claude spawns parallel sub-agents,
# synthesizes, and writes an HTML report into
# your folder. Total: ~2–3 minutes.
Then explore the other 14
/peo — People (team, culture, talent)
/pre — Prediction (experiments, forecasts, premortem)
/play — Playbook (mission, rhythm, postsuccess)
/rev — Revenue (freemium, unit economics)
/rea — Reach (funnels, acquisition)
/ret — Retention (love machine, community)
Want the long-form walkthrough? The full guide on GitHub covers troubleshooting, trigger-phrase invocation, team sharing, manual install for corporate-firewall setups, and Path B for running skills in Claude.ai web chat without installing anything.
Read the full install guide →

Common questions

A HAAO is an organization where AI agents are the default operators for growth, retention, and operational loops — and humans retain authority for strategic decisions only. The agents surface choices; you make the calls. It's the organizational model built for the AI era: autonomous by default, human-assisted at the decision layer.
You need Claude Code installed and basic terminal comfort (copy-paste install). The agents are markdown-based slash commands — no programming required to run them. If you can paste a command and follow prompts, you can use Atomic Scaling OS.
Solo founders and small teams at $0–$10M who want to scale revenue without scaling headcount. Technical founders who want AI handling the operational grind so they can focus on product. Growth operators who need structured frameworks instead of guessing. Anyone who read Atomic Scaling and wants to operationalize the 3P3R Method® with AI.
Atomic Scaling OS is the open-source implementation of the 3P3R Method® from Atomic Scaling: How Small Teams Create Huge Growth. The book provides frameworks and case studies; this repo gives you the executable agents. Available at atomicscaling.com.
Yes — MIT license, fork freely. Adapt the agents to your industry, add new ones for your context, modify thresholds and scoring. Pull requests are welcome. Open an issue first for significant changes.

Built from the field, not the whiteboard.

Ludovic Bodin
Ludovic Bodin
3× Entrepreneur · 2× Unicorn Investor · 1× IPO · Founder, BOBIC Generational Wealth · Author, Atomic Scaling
Ludovic has spent years studying and advising the companies that redefined their industries by staying small — Supercell, Valve, ClickFunnels — and distilled their frameworks into the 3P3R Method®. Atomic Scaling OS is the open-source operationalization of those frameworks as Claude agents, built for the HAAO era.
"If a small team from Finland can launch a global phenomenon like Angry Birds and get over 5 billion downloads, then you can do it too. Atomic Scaling is your ultimate playbook for unlocking explosive growth."
Peter Vesterbacka
Former Mighty Eagle at Rovio, Co-Founder of Angry Birds
"We applied Atomic Scaling to build & scale Gigaverse to over $10M in annualised revenue in our first 6 months, with a team of less than 10 people."
Dith
Co-Founder, Gigaverse
"I am happy to see all these learnings combined in one book to help future entrepreneurs, not just in the games industry but all industries."
Ilkka Paananen
Co-founder & CEO, Supercell
"Ludovic Bodin's Atomic Scaling provides an insightful framework that sheds light on how gamers are driving the future of entertainment, technology, and society."
Kevin Lin
Co-founder, Twitch
"Atomic Scaling is a must-read for any entrepreneur looking to take their business to the next level."
Christine Comaford
New York Times Bestselling Author, "Power Your Tribe"
"I love seeing small teams build big things. This book embodies the idea of leverage and provides tactics for leaders looking to do a lot with a little."
Eric Jorgenson
CEO, Scribe Media · Author, "The Almanack of Naval Ravikant"
"Atomic Scaling equips leaders with a framework to sharpen foresight and navigate scaling challenges effectively."
Risto Siilasmaa
Former Chairman, Nokia · Founder, F-Secure
"In Atomic Scaling, Ludovic shows how boosting talent density, team alignment, and efficiency can overcome capital limits and achieve more with less."
Xiang Wang
Former President, Xiaomi
"Atomic Scaling is a game-changing asset that offers effective strategies for achieving exponential growth and success in the digital age."
Alain Crozier
Former CEO, Microsoft China
"Atomic Scaling shares incredible insights to help navigate the complexities of entrepreneurship. A treasure trove of lessons for gaming and non-gaming founders alike."
Joseph Kim
Co-founder, Lila Games & Gamemakers
"A must-have playbook for anyone seeking to redefine the rules of the game. Atomic Scaling offers powerful insights from the gaming industry's unprecedented success."
Verne Harnish
Founder, Entrepreneurs' Organization (EO) · Author, "Scaling Up"
"Atomic Scaling by Ludovic Bodin - I want all the golden nuggets I can to help founders raise less or avoid VCs entirely. This Book has some killer actionable ideas in it."
Chris Tottman
Founding General Partner of NOTION.vc ($1B AUM) · "The Million Dollar ARR Club"
In the field
"We applied Atomic Scaling to build & scale Gigaverse to over $10M in annualised revenue in our first 6 months, with a team of less than 10 people."
Dith
Co-Founder, Gigaverse · Atomic Scaling advisory
ATOMIC SCALING IN PRODUCTION
Gigaverse scaled to $10M+ annualised revenue in 6 months with fewer than 10 people — applying the 3P3R Method® to maximize output per person across product, growth, and operations.

<10 people · $10M+ ARR in 6 months