Lead Hand

What It Does

Lead Hand is an autonomous lead generation engine that discovers prospects matching your Ideal Customer Profile (ICP), enriches them with web research, scores them 0-100, deduplicates against your existing database, and delivers qualified leads in CSV/JSON/Markdown format—on a schedule. This is not a one-time scrape. Lead Hand runs daily (or weekly) and continuously builds your lead database over time.

Key Features

ICP-based discovery: Targets specific industries, roles, company sizes, and geographies
Multi-query web research: 5-10 search strategies per run to find prospects
Enrichment: Company size, tech stack, funding, recent news
Scoring: 0-100 lead quality score based on ICP match, growth signals, and data completeness
Deduplication: Never delivers the same lead twice
Scheduled delivery: Daily at 7/8/9 AM, weekdays only, or weekly

Activation

# Activate the Lead Hand
openfang hand activate lead

# Activate with custom ICP
openfang hand activate lead --settings "target_industry=SaaS,target_role=CTO,company_size=startup"

# Check status
openfang hand status lead

Configuration Settings

target_industry

text

Industry vertical to focus on (e.g., SaaS, fintech, healthcare, e-commerce)

target_role

text

Decision-maker titles to target (e.g., CTO, VP Engineering, Head of Product)

company_size

select

default:"any"

Filter leads by company size:

any: Any size (default)
startup: 1-50 employees
smb: 50-500 employees
enterprise: 500+ employees

lead_source

select

default:"web_search"

Primary method for discovering leads:

web_search: Web search (default)
linkedin_public: LinkedIn public profiles
crunchbase: Crunchbase data
custom: Specify in prompt

output_format

select

default:"csv"

Report delivery format:

csv: CSV spreadsheet (default)
json: JSON data
markdown_table: Markdown table

leads_per_report

select

default:"25"

Number of leads to include in each report:

10 leads
25 leads (default)
50 leads
100 leads

delivery_schedule

select

default:"daily_9am"

When to generate and deliver lead reports:

daily_7am: Daily at 7 AM
daily_9am: Daily at 9 AM (default)
weekdays_8am: Weekdays at 8 AM
weekly_monday: Weekly on Monday

geo_focus

text

Geographic region to prioritize (e.g., US, Europe, APAC, global)

enrichment_depth

select

default:"standard"

How much context to gather per lead:

basic: Name, title, company
standard: + company size, industry, tech stack (default)
deep: + funding, recent news, social profiles

Required Tools

Lead Hand requires access to these tools (all built-in):

shell_exec — Platform detection
file_read, file_write, file_list — Database and report files
web_fetch, web_search — Lead discovery and enrichment
memory_store, memory_recall — State persistence
schedule_create, schedule_list, schedule_delete — Daily/weekly scheduling
knowledge_add_entity, knowledge_add_relation, knowledge_query — ICP tracking

System Prompt Overview

Lead Hand operates in 7 phases:

Platform Detection

Detects OS, loads lead database, checks if this is a scheduled run or manual trigger.

Target Profile Construction

Builds an Ideal Customer Profile (ICP) from settings: industry, role, company size, geography. Stores ICP in knowledge graph.

Lead Discovery

Executes 5-10 search queries combining industry + role + growth signals (hiring, funding, news). Aims to discover 2-3x the target number to allow for filtering.

Lead Enrichment

For each discovered lead, fetches additional data based on enrichment depth: company website, employee count, tech stack, funding news, social profiles.

Deduplication & Scoring

Compares new leads against existing database, skips duplicates, scores each lead 0-100 based on ICP match, growth signals, enrichment completeness, recency, and accessibility.

Report Generation

Generates report in configured format (CSV/JSON/Markdown), saves to lead_report_YYYY-MM-DD.{csv,json,md}.

State Persistence

Updates leads_database.json with all known leads (new + existing), saves state to memory, updates dashboard metrics.

Usage Examples

Daily Lead Generation

# Configure ICP and schedule
openfang hand configure lead \
  --set target_industry="fintech" \
  --set target_role="Head of Engineering,VP Engineering" \
  --set company_size="smb" \
  --set geo_focus="US" \
  --set delivery_schedule="daily_9am"

# Activate
openfang hand activate lead

Lead Hand will now run every day at 9 AM, discover fintech companies with 50-500 employees in the US, find their engineering leaders, and deliver a report.

Manual Lead Run

# Run immediately without waiting for schedule
openfang hand run lead

Deep Enrichment for High-Value Leads

openfang hand configure lead \
  --set enrichment_depth="deep" \
  --set leads_per_report="10"

openfang hand run lead

Lead Hand will fetch funding news, recent company activity, and social profiles for each lead. With deep enrichment, it delivers fewer leads per run but with much richer context.

Lead Scoring Framework

Each lead is scored 0-100 based on five factors:

ICP Match (30 points max)

Industry match: +10
Company size match: +5
Geography match: +5
Role/title match: +10

Growth Signals (20 points max)

Recent funding: +8
Actively hiring: +6
Product launch: +3
Press coverage: +3

Enrichment Quality (20 points max)

Email found: +5
LinkedIn found: +5
Full company data: +5
Tech stack known: +5

Recency (15 points max)

Active this month: +15
Active this quarter: +10
Active this year: +5
No recent activity: +0

Accessibility (15 points max)

Direct contact info: +15
Company contact: +10
Social only: +5
No contact info: +0

Score Interpretation:

80-100 (A): Hot lead — prioritize outreach
60-79 (B): Warm lead — nurture
40-59 (C): Cool lead — enrich further
0-39 (D): Cold lead — deprioritize

Dashboard Metrics

Lead Hand tracks four key metrics:

Leads Found

Total unique leads discovered across all runs.

Reports Generated

Total lead reports delivered.

Last Report

Date and time of the most recent report.

Unique Companies

Total unique companies in the database.

View in the dashboard at http://localhost:4200/hands/lead.

Output Formats

CSV Format

Name,Title,Company,Company URL,LinkedIn,Industry,Size,Score,Discovered,Notes
"Jane Smith","VP Engineering","Acme Corp","https://acme.com","https://linkedin.com/in/janesmith","SaaS","SMB (120 employees)",85,"2025-01-15","Series B funded, hiring 5 engineers"

JSON Format

[
  {
    "name": "Jane Smith",
    "title": "VP Engineering",
    "company": "Acme Corp",
    "company_url": "https://acme.com",
    "linkedin": "https://linkedin.com/in/janesmith",
    "industry": "SaaS",
    "company_size": "SMB",
    "employee_count": 120,
    "score": 85,
    "discovered": "2025-01-15",
    "enrichment": {
      "funding": "Series B, $15M",
      "hiring": true,
      "tech_stack": ["React", "Python", "AWS"],
      "recent_news": "Launched enterprise plan Q4 2024"
    },
    "notes": "Strong ICP match, actively growing"
  }
]

Markdown Table Format

| # | Name | Title | Company | Score | Key Signal |
|---|------|-------|---------|-------|------------|
| 1 | Jane Smith | VP Engineering | Acme Corp | 85 | Series B funded, hiring |
| 2 | John Doe | CTO | Beta Inc | 72 | Product launch Q1 2025 |

Best Practices

Lead Hand uses only publicly available information. It will never scrape behind login walls or paywalls. All data must be ethically sourced.

Start with a narrow ICP (specific industry + role + size). Once you validate the quality, broaden the criteria.

Check your leads_database.json regularly. Lead Hand never forgets a lead—if you want to reset, delete this file.

Use deep enrichment for high-value enterprise leads. Use basic enrichment for volume plays.

Advanced Configuration

Custom Search Queries

You can manually trigger Lead Hand with custom search guidance:

openfang chat lead
> "Find CTOs at Series A fintech startups in NYC that recently launched a product."

Lead Hand will adjust its search queries to match this specific request.

Integration with CRM

Export leads to your CRM:

# Generate JSON format
openfang hand configure lead --set output_format=json

# Run and pipe to your CRM's API
cat lead_report_2025-03-07.json | curl -X POST https://your-crm.com/api/leads \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -d @-

Deduplication Strategy

Lead Hand uses a three-level matching algorithm:

Exact match: Normalized company name + person name
Fuzzy match: Levenshtein distance < 2 on company name + same person
Domain match: Same company website domain = same company

When a duplicate is found, Lead Hand merges the enrichment data (keeps the most complete record).

Example Output

# Lead Report: 2025-03-07
**ICP**: SaaS | VP Engineering | SMB | US
**Leads Delivered**: 25 | **Average Score**: 68

## Top 5 Leads

1. **Jane Smith** (85) — VP Engineering at Acme Corp
   - Company: 120 employees, Series B funded ($15M)
   - Signal: Hiring 5 engineers, launched enterprise plan Q4 2024
   - LinkedIn: https://linkedin.com/in/janesmith

2. **John Doe** (72) — CTO at Beta Inc
   - Company: 80 employees, Series A funded
   - Signal: Product launch Q1 2025
   - LinkedIn: https://linkedin.com/in/johndoe

[Full CSV attached: lead_report_2025-03-07.csv]

Get Started

Core Concepts

Autonomous Hands

Configuration

Guides

Security

Deployment

What It Does

Key Features

Activation

Configuration Settings

Required Tools

System Prompt Overview

Usage Examples

Daily Lead Generation

Manual Lead Run

Deep Enrichment for High-Value Leads

Lead Scoring Framework

Dashboard Metrics

Leads Found

Reports Generated

Last Report

Unique Companies

Output Formats

CSV Format

JSON Format

Markdown Table Format

Best Practices

Advanced Configuration

Custom Search Queries

Integration with CRM

Deduplication Strategy

Example Output

Next Steps

Collector Hand

Researcher Hand

Build docs developers (and LLMs) love

Get Started

Core Concepts

Autonomous Hands

Configuration

Guides

Security

Deployment

​What It Does

​Key Features

​Activation

​Configuration Settings

​Required Tools

​System Prompt Overview

​Usage Examples

​Daily Lead Generation

​Manual Lead Run

​Deep Enrichment for High-Value Leads

​Lead Scoring Framework

​Dashboard Metrics

Leads Found

Reports Generated

Last Report

Unique Companies

​Output Formats

​CSV Format

​JSON Format

​Markdown Table Format

​Best Practices

​Advanced Configuration

​Custom Search Queries

​Integration with CRM

​Deduplication Strategy

​Example Output

​Next Steps

Collector Hand

Researcher Hand

Build docs developers (and LLMs) love

What It Does

Key Features

Activation

Configuration Settings

Required Tools

System Prompt Overview

Usage Examples

Daily Lead Generation

Manual Lead Run

Deep Enrichment for High-Value Leads

Lead Scoring Framework

Dashboard Metrics

Output Formats

CSV Format

JSON Format

Markdown Table Format

Best Practices

Advanced Configuration

Custom Search Queries

Integration with CRM

Deduplication Strategy

Example Output

Next Steps