"Show up prepared, ask better questions, make real connections."
A Claude skill that helps you research people before meetings so you can have more meaningful conversations. Aggregates public information from LinkedIn, personal websites, academic profiles, and more into warm, readable reports that help you connect authentically β not awkwardly.
Bello helps you prepare for conversations by researching people from publicly available sources. It's like having a thoughtful friend who's an excellent researcher, helping you show up informed and ready to connect.
- π€ Coffee chats & networking β Show up knowing their interests, not just their job title
- πΌ Job interviews β Ask informed questions about their team, company, or recent work
- π° Investor meetings β Understand their portfolio, thesis, and what they actually fund
- π Academic connections β Reference their research, lab culture, and collaboration style
- π’ Professional introductions β Break the ice with genuine shared interests
- Stand out: Most people show up with generic questions. You'll reference their actual work.
- Find common ground: Discover shared interests before the meeting (rock climbing, travel, food)
- Ask thoughtful questions: Reference their recent posts, portfolio companies, or research
- Build authentic rapport: Move past small talk to topics they genuinely care about
- Personal life matters β Hobbies, travel, pets, food interests β the stuff that actually starts conversations
- Finance Intelligence Module β For VC/investor meetings: portfolio analysis, fund strategy, deal patterns
- Specific conversation starters β 6 openers backed by evidence from their actual work
- Multi-source synthesis β LinkedIn + personal websites + academic pages + writing = complete picture
- Warm, human writing β Reads like a friend briefing you, not a background check report
LinkedIn MCP Server (Required for Full Functionality)
Bello works without connectors but gets only ~30% of LinkedIn profiles from search snippets. For full reports (95%+ coverage), install the LinkedIn MCP server:
# Install the LinkedIn MCP server
npm install -g @modelcontextprotocol/server-linkedinThen add to ~/.claude/config.json:
{
"mcpServers": {
"linkedin": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-linkedin"],
"env": {
"LINKEDIN_USERNAME": "your-email@example.com",
"LINKEDIN_PASSWORD": "your-password"
}
}
}
}Security note: Store credentials securely. For OAuth-based options (more secure), see docs/connector-setup.md.
- Clone this repository:
git clone https://github.com/evelynchen916-coder/bello-skill.git
cd bello-skill- Copy the skill to your skills directory:
mkdir -p ~/.claude/skills/bello
cp SKILL.md ~/.claude/skills/bello/SKILL.md- Verify installation:
# In Claude Code
/skills listYou should see bello in the list.
- Test it out:
Tell me about [Person Name at Company]
- Download
SKILL.mdfrom this repository - Copy to
~/.claude/skills/bello/SKILL.md - Restart Claude Desktop
- The skill will automatically activate when you ask to research someone
Research John Smith, VP of Engineering at Anthropic
Prepare me for a coffee chat with Sarah Chen, climate VC at Sequoia
Who is Dr. Emily Rodriguez? She's a computational biology professor at Stanford.
Tell me about Ben Scruggs at Hatteras Venture Partners before our investor meeting
Bello automatically detects finance professionals and activates the Finance Intelligence Module, which includes:
- Portfolio company financials
- Fund economics and deployment strategy
- Investment sector focus with normalized allocations
- Deal-specific conversation starters
Every Bello report includes:
- Identity signals β Quick context (location trail, platforms, confidence level)
- The Pull Quote β Most revealing statement from their own words
- Who They Really Are β 3-5 paragraph portrait with communication style analysis
- What They Think About β Interest/capital allocation table (always sums to 100%)
- Their Journey β Career timeline with insights
- Signal Analysis β Communication style mapping
- Life Outside Work β Hobbies, travel, pets, personal interests
- Conversation Starters β 6 specific openers backed by evidence
- Source Inventory β Transparent accounting of data sources
- What We Didn't Find β Honest gaps in research
- Capital Allocation Table β Normalized to 100% (thesis areas, deployment mechanisms)
- Investment Sector Focus β Stage, check size, portfolio percentage by category
- Fund & Deployment Strategy β AUM, fundraising history, typical check sizes
- Portfolio Company Financials β Total raised, exit status, return signals
- Financial Health Signals β Customer concentration, partnership status, exit velocity
- Deal-Specific Conversation Starters β Reference actual funds, portfolio companies, check sizes
Bello is read-only and uses only publicly available information:
- β Institutional pages, conference bios, academic databases
- β LinkedIn/Twitter search snippets (no login required)
- β Public company websites, press releases, news
- β Personal websites, Substacks, Medium, GitHub
- β Private accounts, locked profiles, non-public data
- β Never automates connection requests, messages, or engagement
Rate limiting: LinkedIn allows ~80-100 profile views/day on standard accounts. Bello typically needs 1-5 views per report.
Bello works without any connectors, but LinkedIn MCP access improves report quality by 3-4x.
Without connector: ~30% of LinkedIn profile from search snippets With connector: Full profile, all posts, experience details, skills, recommendations
- Navigate to the LinkedIn profile in your browser
- Ask Claude to read the page directly
- Works with your authenticated session β no setup required
See docs/connector-setup.md for detailed instructions on:
- ConnectSafely MCP (OAuth, no password sharing)
- LinkupAPI (no-code, free tier)
- Composio LinkedIn integration (for Claude Code)
Bello runs a 6-phase research process:
- Intake β Extract name, context, hints from user query
- Discovery β 15-25 targeted searches across multiple source types
- Extraction β Collect professional, academic, personal, finance data
- Analysis β Pattern detection, network inference, deduction
- Finance Module β (If applicable) Deep dive into portfolio, fund economics, deal strategy
- Writing β Synthesize into magazine-quality report
- Institutional pages (lab pages, company "About" pages) β fully accessible
- Conference bios, speaker pages β fully accessible
- Personal websites, Substacks, Medium, GitHub β fully accessible
- Academic databases (Google Scholar, Semantic Scholar) β accessible
- LinkedIn search snippets β partial access, ~30% of profile
- Finance databases (Crunchbase, PitchBook snippets) β accessible
- News, podcasts, press β fully accessible
Scenario: You have a coffee chat with a senior engineer at a company you admire.
Bello delivers:
- Career trajectory and technical skills
- Recent talks/blog posts to reference
- Hobbies and personal interests for rapport
- Specific conversation starters backed by evidence
Scenario: You're pitching a VC and want to understand their investment thesis.
Bello delivers (Finance Module):
- Portfolio company analysis with financials
- Check size ranges by stage
- Investment sector focus (normalized allocation)
- Co-investment patterns and syndicate strategy
- Deal-specific conversation starters
Scenario: You're considering a PhD advisor or research collaboration.
Bello delivers:
- Publication history and citation patterns
- Advisor relationships and co-author network
- Research focus evolution over time
- Lab culture signals from student/postdoc pages
Scenario: You're interviewing at a startup and want to research the founders.
Bello delivers:
- Founding team analysis (technical backgrounds, complementary skills)
- Company product/technology deep dive
- Communication style and intellectual positioning
- Personal interests for authentic connection
Automatically activates for VCs, private equity, angels, asset managers. Includes:
1. Normalized Capital Allocation
| Focus Area | % of Capital/Time | Trend | Deployment Mechanism |
|-----------|------------------|-------|----------------------|
| NAM infrastructure | 35% | ββ | Lead seed checks, board seats |
| Biopharmaceutical therapeutics | 40% | β | Series A-B follow-ons |
| Regulatory policy & FDA alignment | 15% | β | SPV co-investments |
| Life science tools | 10% | β | Opportunistic |
| **TOTAL** | **100%** | | |2. Portfolio Company Financials
| Company | Stage | Total Raised | Exit Status | Timeline | Return Signal |
|---------|-------|--------------|------------|----------|--------------|
| Kymera Therapeutics | Public | $500M+ | $KYMR | 8 years | >$3B market cap |
| HistoSonics | Acquired | $20M | Bezos Expeditions | 15 years | 112x return |3. Investment Pattern Analysis
- Check sizes by stage (seed: $500K-$2M, Series A: $2M-$8M)
- Follow-on participation rate
- Board seat allocation (reveals conviction)
- Exit velocity and return patterns
Bello runs 15-25 searches across 5-6 different source types:
- Core identity (LinkedIn snippets, institutional affiliations)
- Institutional traces (lab pages, university profiles, conference bios)
- Academic depth (Google Scholar, papers, co-authors)
- Writing & personal (Substack, Medium, podcasts, hobbies)
- Startup/company (Crunchbase, product descriptions, co-founder analysis)
- Finance (portfolio searches, fund data, board seats)
Cross-references across sources to ensure correct person:
- Same university + graduation year β high confidence
- Co-author with known advisor β very high confidence
- Shared surname + complementary skills β flag as potential family connection
Contributions welcome! See CONTRIBUTING.md for guidelines.
- Additional finance database integrations (PitchBook, Dealroom)
- International source support (non-English profiles)
- Academic paper sentiment analysis (citation context)
- Company financial health scoring
- Network graph visualization
MIT License - see LICENSE for details.
Built for Claude Code by Evelyn Chen.
Inspired by the need for authentic, evidence-based networking prep that respects privacy while delivering genuine insights.
- Issues: GitHub Issues
- Discussions: GitHub Discussions
- Email: [Your contact email if you want to include]