Turn a founder's idea into a realistic, investor-credible business plan — full narrative plan, a one-page summary, and a ~300-word investor pitch — as editable Markdown. Its defining feature is honesty. Part of claude-skills.
From your idea, purpose, and stage it builds a full plan: exec summary, problem, solution, market (bottoms-up TAM/SAM/SOM), competitor teardown, business model + pricing options, go-to-market (a concrete "first 10 paying customers" plan), a 12-month financial model with unit economics, risks (likelihood × impact × mitigation), a 90-day action plan, an assumptions register, a validation-gaps checklist, and a straight go / no-go / reshape verdict. Emphasis adapts to purpose — investor (traction, market, the ask), internal (execution, milestones), or loan (cash-flow, break-even, repayment).
The whole point is that every number is trustworthy: market/competitor facts are researched and cited (with an as-of date) or asked or left as a visible placeholder; financials are computed from your assumptions by a deterministic helper (scripts/financials.py) and labeled "derived from your inputs."
- State the purpose — investor vs internal vs loan changes what it foregrounds.
- Bring what you know — your price point, target customer, stage, any competitors or traction. The more real facts you supply, the fewer placeholders.
- Supply financial assumptions for real numbers — price, expected growth/new customers, churn, unit cost, fixed costs, CAC. It runs these through the model; without them it uses clearly-labeled draft assumptions you can replace.
- Trust the verdict — if it says "too crowded" or "the unit economics don't work yet," that's the feature. A plan that only cheerleads isn't useful.
- Close the validation gaps it lists before you show the plan to anyone or spend real money.
The honesty boundary is the point — it will not manufacture credibility:
- It never invents a market size, a competitor's pricing/positioning, or a revenue figure. Unknown facts are researched-and-cited, asked, or marked
[… — validate]— never guessed. - It refuses the "make it look impressive" asks — fill in the TAM, invent competitor numbers, produce a 5-year revenue forecast — and explains why fabricating those makes a plan worthless to an investor or bank.
- Financials are computed from your assumptions, not forecast. It does the arithmetic on your inputs; it does not predict the future or guarantee outcomes.
- It is not legal, tax, or investment advice, and does not guarantee funding. It structures and pressure-tests a plan; the facts and the decision are yours to verify.
uv (for the stdlib financials helper — no runtime dependencies; scripts/financials.py runs via uv run). For market/competitor research it uses web search if available; where research fails, it falls back to labeled placeholders rather than inventing. Optional: report-builder to render a shareable HTML version.