VC Fundraising Skills for Startups

A Claude Code skill that helps founders research and target venture capital investors. Describe your startup and it walks you through finding aligned VC funds, mapping the right people, enriching contact details, scoring fit, and drafting personalized outreach — step by step.

Getting Started

1. Get API keys for Apify and Exa.
2. Export APIFY_TOKEN and EXA_API_KEY in your shell.
3. Open this repo in Claude Code (cd vc-fundraising-skill && claude).
4. Start with: "Start vc fundraising" and follow each review checkpoint.

Install this skills repo

skills.sh and Claude Code plugin manager are not available in this environment yet, so install manually for now.

Manual install (current)

mkdir -p ~/.claude/skills
cp -R .claude/skills/* ~/.claude/skills/

This installs the skills globally under ~/.claude/skills/ so they are available across projects.

To update later, pull the latest changes in this repo and run the copy command again.

When skills CLI / plugins are available (future)

Via skills.sh CLI:

npx skills add extruct-ai/gtm-skills

Via Claude Code plugin manager:

/plugin marketplace add extruct-ai/gtm-skills
/plugin install gtm-skills

Updating:

npx skills add extruct-ai/gtm-skills
# or
npx skills update

Plugin manager update:

/plugin marketplace update

What it does

clarify-thesis → find-vcs → map-investors → enrich-vc-contacts → score-vc-fit → draft-investor-outreach

Step	What happens	You review
1. clarify-thesis	Asks targeted questions to build a structured fundraising brief	Confirm the brief before continuing
2. find-vcs	Searches VC databases (OpenVC, Crunchbase, etc.) and scrapes firm websites	Approve / edit the list of 15–30 candidate firms
3. map-investors	Finds the right partners and principals at each firm	Approve / edit the people list
4. enrich-vc-contacts	Finds public emails and LinkedIn profiles	Review contact quality tiers
5. score-vc-fit	Scores each firm on 8 dimensions and ranks the list	Confirm or reorder the ranking
6. draft-investor-outreach	Drafts a tailored email + LinkedIn message for each investor	Edit and send

Requirements

Two API tokens are required before running the pipeline:

Service	Used for	Get it
Apify	Scraping VC firm websites, team pages, portfolio pages, LinkedIn profiles, and bulk email finding	apify.com → Console → Settings → API tokens
Exa	Semantic people search — finding the right partners and principals at each firm	exa.ai → Dashboard → API keys

Both tokens must be set before opening the repo in Claude Code. The pipeline will not work without them.

Setup

1. Get your API tokens

• Apify: sign up at apify.com, go to Console → Settings → API tokens
• Exa: sign up at exa.ai, go to Dashboard → API keys

2. Set your tokens

export APIFY_TOKEN=your_apify_token_here
export EXA_API_KEY=your_exa_api_key_here

Add both lines to your .bashrc or .zshrc to persist them across sessions:

echo 'export APIFY_TOKEN=your_apify_token_here' >> ~/.zshrc
echo 'export EXA_API_KEY=your_exa_api_key_here' >> ~/.zshrc
source ~/.zshrc

3. Open the repo in Claude Code

cd vc-fundraising-skill
claude

Claude Code will automatically pick up the Apify and Exa MCP servers from .mcp.json.

4. Start the pipeline

Say:

"I'm fundraising and want to find relevant VCs. My company does [brief description]."

Or simply:

"Start vc fundraising"

The pipeline will guide you through each step, pause for your review at every checkpoint, and save structured outputs to data/ as you go.

Outputs

All outputs are saved in data/:

File	Contents	Step
data/normalized/company_profile.json	Structured fundraising brief	After step 1
data/raw/fund_candidates.json	Candidate VC firms with rationale	After step 2
data/raw/investor_list.json	Individual investors at each firm	After step 3
data/normalized/investor_contacts.json	Enriched contacts with emails + LinkedIn	After step 4
data/normalized/scored_funds.json	Scored and ranked firms with breakdowns	After step 5
data/normalized/final_output.json	Ranked outreach list with drafted messages	After step 6

The final_output.json file contains everything you need to send outreach — investor name, contact details, subject line, email body, and LinkedIn message.

Notes

• Only public professional contact information is used
• Inferred emails are clearly labeled as contact_tier: "inferred" — verify before sending
• Investors with no email are still included for LinkedIn outreach
• Each step waits for your confirmation before proceeding
• You can re-run any step individually without restarting the whole pipeline

---

Built by Overdrive