Stress-Test a Business Idea Before Launch With 11 Founder Skills

Most business ideas die after the money is spent, not before. The Founder skill is a 2k-star set of eleven Claude skills that test a business before you launch it — free, MIT, no signup, no API key, nothing to connect.

Why This Skill Matters

The eleven skills divide the work of a founding team. One is a board of directors whose three members each apply one published framework: Offers from Alex Hormozi's $100M Offers, Monopoly from Peter Thiel's Zero to One, and Product from Walter Isaacson's biography of Steve Jobs. One is a CFO that finds your real profit margins. One is a consumer panel that spins up buyer agents (100 by default) trained on your target customer and tallies who buys, who passes, and why.

Each skill reads what the others wrote in a founder/ folder in your project, so the pricing strategist works from the panel's answers and the operations manager feeds real costs back to the CFO.

Installation

The README's first path is to paste the repo link into Claude and say install skill. As a Claude Code plugin:

/plugin marketplace add Jakeschincariol/founder-skill
/plugin install founder-skill@founder-skill

Claude Code namespaces plugin skills, so they appear as /founder-skill:founder-board and so on. For plain /founder-board, copy the folders instead:

git clone https://github.com/Jakeschincariol/founder-skill.git
cp -r founder-skill/skills/founder-* ~/.claude/skills/

Copy the same folders into your repo's .claude/skills/ for a project-local install. Without Claude Code at all, you can paste any single SKILL.md at the top of a chat and it runs as a mode — you lose the sub-agents and the Python tools, but the method works. The tools need Python 3.8 or newer with nothing to pip install.

Real Workflow: Test a Matcha Café

The README's worked example is a matcha café in Toronto selling iced matcha lattes at $6.50 a cup. Every input ships in the repo's example files, so you can run the tools on it before trying your own idea. Run the skills in this order:

board -> competitors -> consumer -> pricing -> offer -> cfo -> marketing -> brand -> ops -> launch -> plan

1. /founder-board writes founder/idea.md and puts it in front of the three board members. Each scores it, lists the risks, and votes: fund, fund if, or pass. The conditions become a checklist for the rest.
2. /founder-competitors maps the matcha bars and coffee shops within walking distance, what they charge, and what their reviews complain about — every fact linked to a public source.
3. /founder-consumer builds the target customer (urban professionals, 24 to 38, who buy coffee most days), deals 100 buyer cards across income, habits, and objections, and runs one sub-agent per buyer. Back come who buys, who passes, the reasons in the buyers' words — "my coffee place is closer", "$6.50 is steep for a drink" — and what would flip a no.
4. /founder-pricing reads the panel's four price answers per buyer into an acceptable range and checks $6.50 against it and the competitors.
5. /founder-offer answers the top objections (a launch-week price, a loyalty card) and re-runs a quick panel to see if the buy rate moves.
6. /founder-cfo runs the unit economics. With the example numbers, each cup leaves $4.76 after its own costs, break-even is 186 cups a day, and the $25,280 startup spend is earned back in month 8.
7. /founder-marketing, /founder-brand, /founder-ops, and /founder-launch turn it into a 30-day campaign, a name and a look, the suppliers and routines, and a dated launch that starts with a pop-up to test real buyers.
8. /founder-plan compiles founder/business-plan.md. Its verdict is computed, not written: Profitable, Not yet, or Incomplete.

Tips

• Any skill runs on its own. /founder-consumer --quick on an idea you are only thinking about is the README's suggested way to find out if you should keep thinking.
• The four Python tools (unit_economics.py, panel.py, van_westendorp.py, compile.py) are standard-library only and never touch the network. The CFO tool treats a missing field as an error, never a guess.
• Read the panel's buy rate as an upper bound. The README notes that language models lean agreeable, and /founder-launch starts with a cheap real-world test for exactly that reason.
• Panel answers are research, never testimonials. The same fine print says the board applies published frameworks in its own words — it never speaks as Hormozi, Thiel, or Jobs, and the project is not affiliated with any of them.

When Not to Use This

This is not financial, legal, or tax advice: the README says permits, trademarks, and claims depend on where you are, and that a professional confirms them. The numbers are only as good as your inputs, so the CFO marks every estimate as an estimate and asks for real quotes. And the consumer panel is simulated buyers, not customers — confirm the big calls with real people before you commit money.

See the leaderboard for more skills.