Skip to main content
skill CX Codex

yc-advisor — Claude Code skill: what it does, how to install

A skill that connects AI assistants to 443 curated Y Combinator resources: Paul Graham essays, partner lectures, and founder interviews for startup advice.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add agent37-platform/agent37-skills-collection --skill yc-advisor -a claude-code
Open on GitHub

What yc-advisor does

The skill gives the assistant direct access to 443 Y Combinator startup resources. It covers founder equity splits, MVP development, SAFE fundraising notes, and enterprise sales tactics. Instead of generating generic business suggestions, the assistant pulls direct citations from original essays and lecture transcripts.

Retrieval runs in stages to preserve context space. The agent first inspects references/quick-index.md, a lightweight ~500-line index organized by topic. For detailed keyword matches, it runs grep against references/summaries.md. Once the relevant documents are found via references/{CODE}-*.md, the agent reads the full source files before answering. Reading references/index.yaml directly is strictly forbidden because its 64K tokens exceed model context budgets.

Pre-built decision frameworks live in references/frameworks/, covering choices like bootstrap-vs-raise.md or when-to-pivot.md. The workflow forces the agent to cite the author and title for every claim.

What it helps with

  • structure seed fundraising terms and SAFE agreements
  • split equity fairly between technical and business co-founders
  • evaluate readiness for product-market fit before scaling
  • prepare an application for the Y Combinator accelerator batch

Example prompts for the agent

  • Review our co-founder equity split options using YC partner essays and frameworks.
  • What does Paul Graham recommend when deciding whether to bootstrap or raise angel funding?
  • Check the YC archive for indicators on when a startup should pivot.

How to install yc-advisor

1Skills CLI (Vercel)

npx skills add agent37-platform/agent37-skills-collection --skill yc-advisor -a claude-code

2Manually: put the files in place

~/.claude/skills/yc-advisor/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/yc-advisor/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about yc-advisor

How to install yc-advisor in Claude Code?

Run npx skills add agent37-platform/agent37-skills-collection --skill yc-advisor -a claude-code. The skill file will be placed in ~/.claude/skills/yc-advisor/SKILL.md.

What does yc-advisor do?

It connects your coding agent to 443 Y Combinator essays, transcripts, and founder interviews. The agent navigates indexed files, reads full source documents, and produces advice backed by specific partner quotes.

Why does the skill avoid reading index.yaml?

The metadata file references/index.yaml consumes 64K tokens, which exceeds runtime limits. The skill searches via references/quick-index.md and reads specific source files individually using glob patterns instead.

Source text

SKILL.md · MIT
---
name: yc-advisor
description: This skill should be used when the user asks questions about startups, founding decisions, co-founders, fundraising, product development, growth, hiring, or any entrepreneurial advice. It provides access to Y Combinator's complete library of 443 curated resources including essays by Paul Graham, founder interviews, and startup school lectures. Use this skill to give thorough, research-backed advice on startup decisions.
---

# YC Advisor

## Overview

This skill provides access to Y Combinator's comprehensive library of 443 startup resources - essays, podcast transcripts, and video transcripts from YC partners, successful founders, and industry experts.

## How to Use This Skill (Tiered Retrieval)

**Key Principle:** Use quick-index for discovery, but ALWAYS load full source content before answering.

### Step 1: Understand Context (for broad questions)

For broad questions, clarify the user's context:
- **Stage:** Pre-idea | Idea | Building MVP | Launched | Scaling
- **Type:** B2B | Consumer | Hardware | AI/ML | Marketplace
- **Role:** Technical founder | Non-technical | Solo | With co-founder(s)

### Step 2: Discovery

1. Load `references/quick-index.md` to scan available resources (~500 lines, grouped by topic)
2. Identify 3-5 most relevant resources based on the question (line counts help estimate size)
3. Check `references/learning-paths.md` if user is on a founder journey
4. Check `references/frameworks/` for decision questions - use glob `references/frameworks/*.md` to list files, then read specific ones
5. For deeper search, use grep on `references/summaries.md` (too large to load fully)

### Step 3: Deep Dive

1. **Find files using the code** - use glob pattern `references/{CODE}-*.md`
   - Example: For code `DZ`, use glob `references/DZ-*.md` to find the file
   - **WARNING:** NEVER read `index.yaml` - it exceeds token limits (64K tokens)
2. Load the FULL content of top 2-3 resources
3. Read completely - do not skim
4. Extract key insights, quotes, and actionable advice

### Step 4: Synthesize Answer

1. Combine insights from multiple sources
2. Quote directly from source material when valuable
3. **Always cite author and title** for each point
4. Acknowledge tradeoffs and contradictions between sources
5. **Never answer from summaries alone** - always load full source content

## Topic Categories

The library covers these main areas (use for initial filtering):

- **Getting Started:** Should you start? Startup ideas, order of operations, student founders
- **Co-founders:** Finding, relationships, equity splitting, technical vs non-technical
- **Product:** MVP, product-market fit, design, building for users
- **Fundraising:** Seed, Series A, investor pitching, SAFEs, term sheets
- **Growth & Metrics:** Growth strategies, KPIs, conversion, retention
- **Customers & Sales:** Talking to users, first customers, pricing, enterprise sales
- **Hiring & Team:** First hires, engineering teams, equity, management
- **Culture & Leadership:** Building culture, CEO evolution, board management
- **Common Mistakes:** Startup killers, financial health, when to quit
- **Pivoting & Launching:** Pivot strategies, launch timing, press
- **Scaling:** Later stage advice, unicorn characteristics
- **Mindset:** Resourcefulness, handling rejection, goal setting
- **AI Startups:** AI opportunity, moats, vertical agents, vibe coding
- **Founder Interviews:** Airbnb, Stripe, Coinbase, Reddit, Twitch, DoorDash
- **Specialized:** Hardware, biotech, dev tools, crypto, location
- **Joining Startups:** Choosing a startup, stages, equity
- **YC Application:** Application tips, process, YC effect
- **Legal:** Startup mechanics, terms, agreements

## Usage Guidelines

### For Complex Decisions

Questions like "Should I start my own startup or co-found with someone?":

1. Load quick-index.md to identify relevant resources
2. Read 3-5 full source files covering different perspectives
3. Synthesize across sources - look for consensus and contradictions
4. Present balanced view acknowledging tradeoffs
5. Cite specific authors and titles
6. Ask clarifying questions about user's specific situation

### For Factual Questions

Questions like "What are the most common mistakes that kill startups?":

1. Use quick-index.md to find the most authoritative source
2. Load and read the full source file
3. Present comprehensively - don't over-summarize
4. Cite the source

### For Learning Journeys

When users want to learn systematically:

1. Check `references/learning-paths.md` for curated sequences
2. Guide them through resources in order
3. Summarize key takeaways at each step

## Resources

### references/quick-index.md (Primary Discovery)
Lightweight index (~500 lines) grouped by topic. Each entry shows:
- Code, title, author, type, line count, founder stage
- **Use this first** - small enough to load fully
- Use glob pattern `references/{CODE}-*.md` to find files by code

### references/summaries.md (Deep Search)
Detailed summaries with content previews (~4300 lines). Too large to load fully.
- Use grep to search for specific keywords
- Provides more context than quick-index when needed

### references/index.yaml (Maintenance Only - DO NOT READ)
Structured metadata for all resources. **Too large for runtime use (64K tokens).**
Used only by maintenance scripts. For filename lookups, use quick-index.md instead.

### references/learning-paths.md
Curated resource sequences for common founder journeys:
- First-time founder path
- AI startup path
- Fundraising path
- And more...

### references/frameworks/ (Use glob to list, NOT Read)
Decision frameworks for common questions. **Use glob `references/frameworks/*.md` to list files.**
Available frameworks:
- should-i-start-a-startup.md
- solo-vs-cofounder.md
- bootstrap-vs-raise.md
- when-to-pivot.md
- when-to-quit.md
- technical-cofounder-needed.md

### references/*.md
The 443 full-content source files. Each follows this structure:

```markdown
# [Title]

**Author:** [Author Name]
**Type:** [Essay|Podcast|Video]
**URL:** https://www.ycombinator.com/library/[CODE]-[slug]

---

[Full content - essays, transcripts]
```

File naming: `[CODE]-[descriptive-name].md` (e.g., `8z-how-to-get-startup-ideas.md`)

© agent37-platform · MIT · agent37-platform/agent37-skills-collection

Similar skills

All skills for Finance and business →

From the same set2

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor