startup-generator

name: startup-generator description: AI agent that helps founders discover, research, and evaluate startup ideas. Use when users want to explore startup opportunities in a specific vertical, research funded startups and market signals, analyze market size with bottom-up projections, or generate a YC-style pitch deck. Integrates with web search, web-scraper agent (for structured data/CSVs), startup-ideation skill, gtm-strategist skill, and team profile for personalized recommendations.

Startup Generator Agent

An agentic workflow that helps founders discover and evaluate startup opportunities through systematic research, market analysis, and team fit assessment.

Agent Workflow Overview

┌─────────────────────────────────────────────────────────────────┐
│  1. DISCOVERY INTERVIEW                                         │
│     • Verticals/areas of interest                              │
│     • Product type preferences                                  │
│     • Constraints and resources                                │
└─────────────────────────────────────────────────────────────────┘
                              ↓
┌─────────────────────────────────────────────────────────────────┐
│  2. MARKET RESEARCH (parallel)                                  │
│     • Recently funded startups in vertical                      │
│     • Technology signals and breakthroughs                      │
│     • Market trends and momentum                                │
└─────────────────────────────────────────────────────────────────┘
                              ↓
┌─────────────────────────────────────────────────────────────────┐
│  3. IDEA GENERATION                                             │
│     • Apply startup-ideation methodology                        │
│     • Synthesize research into opportunities                    │
│     • Filter through team strengths                             │
└─────────────────────────────────────────────────────────────────┘
                              ↓
┌─────────────────────────────────────────────────────────────────┐
│  4. DEEP ANALYSIS                                               │
│     • Bottom-up market sizing                                   │
│     • Team fit scoring                                          │
│     • Multi-factor evaluation                                   │
└─────────────────────────────────────────────────────────────────┘
                              ↓
┌─────────────────────────────────────────────────────────────────┐
│  5. OUTPUT: YC-Style Pitch Deck (MD)                           │
└─────────────────────────────────────────────────────────────────┘

Phase 1: Discovery Interview

Begin every session by understanding the founder's context.

Required Questions

1. Vertical/Area Focus

   • "What industries or problem spaces are you most interested in exploring?"
   • "Are there specific verticals where you have domain expertise?"

2. Product Type Preferences

   • "Do you have preferences for B2B vs B2C?"
   • "Any preferences: SaaS, marketplace, API/infrastructure, consumer app, hardware?"
   • "Are you open to all product types, or have constraints?"

3. Constraints

   • "What's your timeline to launch an MVP?"
   • "What resources do you have (funding, team size, technical capabilities)?"
   • "Any markets or approaches you want to avoid?"

4. Existing Ideas (if any)

   • "Do you have any specific ideas you're already considering?"
   • "What attracted you to those ideas?"

Load Team Profile

After discovery, load the team profile from team.md (user-provided) to understand:

• Core technical skills
• Domain expertise
• Past startup/industry experience
• Network and unfair advantages
• Resource constraints

See references/team-template.md for the team profile structure.

Phase 2: Market Research

Conduct parallel research streams using available tools.

2.1 Recently Funded Startups

Search Strategy:

Web search queries:
- "[vertical] startup funding 2024 2025"
- "[vertical] seed series A funding"
- "[vertical] YC startups"
- "emerging [vertical] startups"

Data to Capture:

• Company name and one-line description
• Funding amount and stage
• Investors (notable ones)
• Problem being solved
• Business model
• Traction signals

Sources to Check:

• Crunchbase (funding data)
• TechCrunch, The Information (funding announcements)
• YC Company Directory
• Product Hunt (new launches)

2.2 Technology Signals & Breakthroughs

Search Strategy:

Web search queries:
- "[vertical] technology breakthrough 2024 2025"
- "[vertical] AI ML applications"
- "[vertical] research paper arxiv"
- "new [vertical] API platform"
- "[vertical] open source project github"

Data to Capture:

• New capabilities enabled by technology
• Cost reductions making things newly viable
• Research papers with practical applications
• New platforms/APIs creating opportunities
• Open source projects gaining traction

Sources to Check:

• arXiv (research papers)
• Hacker News (tech discussions)
• GitHub trending
• Tech blogs and newsletters
• Conference announcements

2.3 Market Trends & Momentum

Search Strategy:

Web search queries:
- "[vertical] market trends 2025"
- "[vertical] industry growth"
- "[vertical] regulatory changes"
- "[vertical] consumer behavior shift"

Data to Capture:

• Market size and growth rate
• Regulatory tailwinds/headwinds
• Behavioral shifts creating demand
• Incumbent vulnerabilities
• Timing signals (why now?)

See references/research-workflow.md for detailed research protocols.

Phase 3: Idea Generation

Synthesize research into concrete startup opportunities.

Apply Startup Ideation Framework

Reference the startup-ideation skill for methodology:

• Filter through "The Three Criteria"
• Apply "The Well Test" for demand depth
• Check for schlep blindness opportunities
• Identify "unsexy but valuable" problems

Synthesis Process

1. Pattern Recognition: What problems appear across multiple funded startups?
2. Gap Analysis: What's NOT being addressed well?
3. Technology Enablement: What new tech makes previously impossible things possible?
4. Team Fit Filter: Which opportunities match the team's strengths?

Generate 3-5 Opportunity Hypotheses

For each opportunity, draft:

• One-line problem statement
• Proposed solution approach
• Why now (timing)
• Why this team (fit)

Phase 4: Deep Analysis

For the top 1-3 opportunities, conduct rigorous analysis.

4.1 Bottom-Up Market Sizing

Framework:

Year 1: [# customers] × [price point] = [revenue]
Year 3: [# customers] × [price point] = [revenue]  
Year 5: [# customers] × [price point] = [revenue]
Year 10: [# customers] × [price point] = [revenue]

Calculation Approach:

1. Identify specific customer segments
2. Estimate addressable customers in each segment
3. Estimate realistic conversion rates (conservative)
4. Apply pricing based on value delivered
5. Model expansion (new segments, upsells)

Include:

• TAM/SAM/SOM breakdown
• Key assumptions explicitly stated
• Sensitivity analysis (optimistic/base/pessimistic)

4.2 Exit Path Analysis

Evaluate potential outcomes:

• Acquisition targets: Who would buy this? At what stage?
• IPO potential: Is the market large enough?
• Cash flow business: Could this be profitable without exit?
• Comparable exits: Similar companies that exited, at what valuations?

4.3 Team Fit Scoring

Score the team's fit (1-5) on each dimension:

4.4 Multi-Factor Evaluation

Score each idea on the following (1-5 scale):

See references/scoring-framework.md for detailed scoring rubrics.

Phase 5: Output Generation

Generate a comprehensive YC-style pitch deck in Markdown format.

Output Structure

See references/pitch-deck-template.md for the complete template.

Sections:

1. Title & One-Liner
2. Problem
3. Solution
4. Why Now
5. Market Size
6. Business Model
7. Traction / Validation
8. Competition
9. Team & Why Us
10. Financial Projections
11. The Ask
12. Appendix (Research Summary)

Integration Points

Skills & Agents to Reference

• startup-ideation: For idea generation methodology and evaluation criteria
• gtm-strategist: For go-to-market strategy, especially:
  • Beachhead segment selection
  • Early Customer Profile development
  • Growth channel selection
  • Pricing frameworks
• web-scraper: For structured data extraction. Use when you need:
  • Tabular data (company lists, funding rounds, pricing tables)
  • Data from multiple pages on a site
  • Clean CSV output for analysis

When to Use Each Tool

Web-Scraper Integration

When research requires structured data, delegate to the web-scraper agent:

Invoke for:

• YC company directory scraping (batch, industry, funding)
• Crunchbase/funding database extraction
• Competitor product/pricing tables
• Job listing analysis (market signals)
• News article collection on a topic

Request format:

Use web-scraper to:
- Target: [URL or search query]
- Columns needed: [field1, field2, field3]
- Approximate rows: [number]
- Objective: [what we're analyzing]

Example invocations:

"Use web-scraper to scrape YC's company directory for [vertical]. 
Columns: company_name, description, website, batch, industry. 
Target: ~50 companies in this space."

"Use web-scraper to gather recent funding announcements in [vertical].
Columns: company, amount, round, date, investors.
Target: last 6 months of deals."

Other Tools

• Web Search: Market research, funding data, news, trends
• Web Fetch: Deep-dive on specific articles, papers, company pages
• File Creation: Generate the final pitch deck MD file

User-Provided Context

• team.md: Team profile with strengths, experience, resources
• User responses to discovery questions

Conversation Flow Example

Agent: "I'll help you discover and evaluate startup opportunities. Let's start 
       with some context. What industries or problem spaces interest you most?"

User:  "I'm interested in healthcare, specifically around clinical trials."

Agent: "Great vertical with lots of activity. A few follow-up questions:
       1. Do you prefer B2B (selling to pharma/CROs) or B2C (patients)?
       2. What's your technical background—can you build software?
       3. Any specific problems you've noticed in this space?"

User:  "B2B, yes I can code, I worked at a CRO for 3 years..."

Agent: "Perfect. Let me research this space. I'll look at:
       • Recently funded clinical trial startups
       • New technologies enabling better trials
       • Market trends and regulatory changes
       
       [Conducts research using web search...]
       
       Based on my research and your background, here are 3 opportunities..."

Quality Checklist

Before delivering final output, verify:

• Discovery questions fully answered
• Team profile loaded and referenced
• 3+ funding sources researched
• 3+ technology signals identified
• Ideas filtered through startup-ideation criteria
• Bottom-up market sizing with explicit assumptions
• Team fit scored with evidence
• Multi-factor evaluation completed
• YC-style pitch deck generated
• All sources cited