ferdinandobons/startup-skill / startup-design
startup-design
VerifiedDesign, validate, and plan a startup from scratch. Covers market research, competitive analysis, business model, brand identity, product definition, financial projections, and validation experiments. Trigger when the user has a startup idea to explore, wants to validate a business concept, needs a business plan or lean canvas, asks for market sizing or competitive landscape, wants brand positioning or go-to-market strategy, or says anything like "I have an idea for..." or "is this idea worth pursuing". Also handles resuming from a previous checkpoint.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 31.8 KB |
| references/customer-interview.md | markdown | 8.6 KB |
| references/frameworks.md | markdown | 6.1 KB |
| references/honesty-protocol.md | markdown | 5.0 KB |
| references/industry-benchmarks.md | markdown | 2.8 KB |
| references/intake-questions.md | markdown | 2.6 KB |
| references/output-guidelines.md | markdown | 6.5 KB |
| references/output-specs.md | markdown | 5.7 KB |
| references/research-principles.md | markdown | 2.7 KB |
| references/research-scaling.md | markdown | 4.8 KB |
| references/research-synthesis.md | markdown | 6.4 KB |
| references/research-wave-1-market.md | markdown | 6.7 KB |
| references/research-wave-2-competitors.md | markdown | 7.5 KB |
| references/research-wave-3-customers.md | markdown | 8.3 KB |
| references/research-wave-4-distribution.md | markdown | 4.1 KB |
| references/verification-agent.md | markdown | 5.1 KB |
references/research-principles.md
View on GitHub# Research Principles
These principles apply to ALL research agents across all waves. Every agent must follow them.
## Iterative Deep Research
Each agent performs **5-8 web searches minimum**, organized in sequential rounds that drill deeper:
- Round 1: Broad overview queries
- Round 2: Drill-down into specific findings from Round 1
- Round 3: Cross-reference and validate
- Round 4: Reality check and edge cases
Do NOT stop after a single query. The first search gives you the surface — the follow-ups give you the insight.
## Source Quality Tiers
Rank every finding by source reliability:
| Tier | Source Type | Use For |
|------|-----------|---------|
| **Tier 1** | Industry reports (Gartner, Forrester, McKinsey, IBISWorld), SEC filings, government data, academic papers | Hard numbers, market size, growth rates |
| **Tier 2** | Reputable tech press (TechCrunch, Bloomberg, WSJ), company press releases, investor presentations | Funding data, company news, expert opinions |
| **Tier 3** | Blog posts, opinion pieces, social media | Sentiment, customer voice, anecdotal evidence |
A Tier 3 source saying "the market is $5B" is not reliable. A Tier 1 report saying the same thing is.
## Cross-Referencing
Never trust a single source for important claims. For every key finding:
- Look for 2-3 independent sources
- If sources agree: note convergence and cite all
- If sources disagree: note both figures, explain the discrepancy, and state which you trust more and why
## Quantification
Vague claims are worthless. Always push for numbers:
- Bad: "The market is growing fast"
- Good: "$4.2B market growing at 12.3% CAGR (Grand View Research, 2025)"
- Bad: "Many companies use this"
- Good: "67% of mid-size companies report using at least one AP automation tool (Ardent Partners, 2025)"
## Dating
Always note when data was published. Flag anything older than 18 months as potentially outdated. Markets can shift significantly in that time, especially in tech.
## Handling Research Failures
Sometimes WebSearch won't find what you need. When this happens:
1. **Try alternative queries.** Rephrase, use synonyms, try different angles. At least 3 variations before declaring a gap.
2. **Use proxy data.** If you can't find market size for your exact segment, find the broader market and estimate the segment share. Show your math.
3. **Declare the gap explicitly.** Write: "DATA GAP: Could not find reliable data on [X]. Closest proxy: [Y]. Confidence: Low."
4. **Never fabricate.** If you don't have data, say so. An honest "unknown" is infinitely more valuable than a made-up number.
5. **Suggest how to fill the gap.** Point to specific reports that might have the data (even if paywalled), or recommend primary research the founder could do.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |