founder-decision-skill

Founder Decision Copilot 🧭

A Claude skill that acts as a sharp, experienced co-founder for the hard calls — pricing, hiring, firing, pivoting, taking investment, launching, killing a feature, quitting the day job, splitting equity.

It doesn't hand you a balanced pros-and-cons list. It stress-tests your thinking like a demanding partner who's on your side: frames the decision, pokes your weakest assumption, runs a pre-mortem, checks the call against how startups actually die, applies distilled lessons from the entrepreneurial canon, and closes with a real recommendation — plus what would change it.

The failure mode of "startup wisdom" tools is regurgitating quotes. Nobody's life improves because you told them "aim for monopoly." The books are the fuel; the product is a decision process applied to your situation. That's what this is.

What's inside

founder-decision/
├── SKILL.md                          # the 6-step process + the "co-founder who argues" posture
└── references/
    ├── mental-models.md              # 12 models, each with the question it forces on you
    ├── failure-patterns.md           # how startups actually die + real post-mortem databases
    ├── book-principles.md            # the book canon distilled into applicable principles
    ├── investors-and-accelerators.md # YC / a16z / Sequoia / Founders Fund / Draper wisdom
    ├── founder-operating-state.md    # the founder as decision instrument (sleep, stress, focus)
    ├── podcasts-and-voices.md        # non-canon audio sources (regional & science voices)
    ├── decision-frameworks.md        # pre-mortem, WRAP, two-way doors, expected value…
    └── your-context.example.md       # optional: anchor the copilot to YOUR venture

The book canon: The Art of War, Zero to One, The Hard Thing About Hard Things, The Lean Startup, The Mom Test, Shoe Dog, Start-up Nation, High Output Management, Blitzscaling, the Isaacson/Vance biographies of Steve Jobs and Elon Musk, The Diary of a CEO, Paul Graham, Naval, The Innovator's Dilemma, Crossing the Chasm — plus a "further canon" list of strong candidates to add next.

Investor & operator wisdom: Y Combinator / Startup School, Andreessen Horowitz, Sequoia, Founders Fund, Draper, First Round — with a built-in caveat that VC advice optimizes for venture-scale outcomes and can mislead bootstrappers.

The founder's operating state: grounded in the science popularized by Huberman Lab — because a decision run on a sleep-deprived, stress-spiked brain is a worse decision, no matter how good the process on top of it.

Real failure post-mortems: grounded in Failory and startups.rip so the "how startups die" checks cite lived stories, not abstractions.

Regional & audio voices: a scaffold for non-canon sources (e.g. Argentina's La Fábrica for real-economy founder stories, Purple Mango for science/biotech ventures) whose episode-level lessons are added from actual listens rather than fabricated — because faithful distillation beats confident invention.

How the copilot thinks

Every decision runs through the same process (depth scaled to the stakes):

1. Is this even the real decision? — often "should I lower my price?" is really "why isn't anyone buying?" It reframes before analyzing.
2. What kind of door is this? — reversible (decide fast) vs. irreversible (slow down).
3. The 2–3 mental models that actually apply — not a dump; the ones that cut.
4. A pre-mortem — "it's a year later and this blew up. Why?"
5. A check against how startups actually die — does this push toward or away from a known grave?
6. The canon, applied with judgment — the idea, not the quote.
7. The call — a real lean, its confidence level, and what would change the answer.

Install

This is a Claude skill. Drop the founder-decision/ folder into your Claude skills directory:

Claude Code (personal skills):

cp -R founder-decision ~/.claude/skills/

Project-scoped (share it with a repo's collaborators):

cp -R founder-decision /path/to/your/project/.claude/skills/

Restart Claude Code (or start a new session) and it'll appear in your available skills. Then just describe a decision you're weighing — you don't have to name the skill. Things like "I'm torn between…", "should I…", "help me decide whether to…" will trigger it.

Make it yours (optional)

The skill is generic on purpose — it works for any venture. To anchor it to your business so the advice lands on your reality:

cd ~/.claude/skills/founder-decision/references
cp your-context.example.md your-context.md
# fill in your market, stage, model, constraints…

The copilot reads your-context.md automatically if it's there. It's git-ignored in this repo so your private context never gets committed or shared.

Extend it

The skill is built to grow. Each reference file is a self-contained, modular knowledge base — adding to it improves every future decision without touching the core process:

• A book that shaped you? Add a section to references/book-principles.md — source, then a few applicable principles (the idea, not a summary), and an "Apply to:" line.
• A mental model you lean on? Add it to references/mental-models.md in the same format: what it is, the question it forces, the failure it prevents.
• A failure you've lived or witnessed? Add the pattern to references/failure-patterns.md.
• A decision procedure? Add it to references/decision-frameworks.md.

Keep entries short and actionable — the whole design principle is "fuel for a process," not a library. Pull requests welcome.

License

MIT — see LICENSE. Share it with your founder friends.