generate-leads — Claude Code command: ICP lead generation
generate-leads — a Claude Code command that builds a lead list from ICP criteria: industry, company size, tech stack, location. Run it with --enrich comprehensive to get a full buying committee map, technographics, and intent signals.
Apache-2.0 ★ 450 updated 2026-04-03
.claude/commands/generate-leads.mdWhat generate-leads does
The command runs in Claude Code under the /sales-prospecting:generate-leads namespace. It takes ICP criteria via --criteria, a lead count (--count, default 20), and an enrichment level (--enrich: basic, standard, or comprehensive). Output goes to CSV, JSON, or directly into Salesforce, HubSpot, Outreach.io, or SalesLoft format.
The process has five stages: parsing criteria into filters, finding and ranking companies by fit score, identifying decision makers and buying influencers, appending firmographic and technographic data, and generating personalized outreach recommendations. At the comprehensive level, this expands to competitive intelligence and an account research brief. Data processing runs on claude-haiku-4-5; insight generation uses claude-sonnet-4.
Each lead record carries an intent_score, trigger events (for example, a new VP Sales hire or a Series B close), the technology stack, and a recommended contact approach. The --minimum-score 70 flag cuts weak matches before export, so the list you get has already been filtered.
What it helps with
- build a company list from specific ICP criteria with fit scoring
- map the buying committee and surface decision-maker contacts
- append technographics and intent signals to each lead record
- export leads to Salesforce or HubSpot without manual reformatting
- exclude existing CRM records and competitor domains from results
Example prompts for the agent
- Generate 50 B2B SaaS leads with full enrichment:
/sales-prospecting:generate-leads --criteria "B2B SaaS" --count 50 --enrich comprehensive - Find financial-services companies with revenue above $100M in the Northeast US, five contacts per account:
/sales-prospecting:generate-leads --criteria "Financial services, >$100M revenue, Northeast US" --count 50 --enrich comprehensive --contacts 5 - Pull recently funded B2B companies with intent signals and export to HubSpot:
/sales-prospecting:generate-leads --criteria "Recently funded B2B companies" --intent true --enrich standard --format hubspot
How to install generate-leads
Manually: put the files in place
.claude/commands/generate-leads.mdCall it with /generate-leads. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/generate-leads.mdCall it with /prompts:generate-leads. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/generate-leads.prompt.mdManually: put the files in place
.gemini/commands/generate-leads.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about generate-leads
How do I install generate-leads in Claude Code?
Copy the command file to .claude/commands/generate-leads.md in your project repository. The command becomes available as /sales-prospecting:generate-leads immediately after.
What does the generate-leads command do?
It searches for companies matching your ICP criteria, identifies contacts within each account, enriches the data with firmographics and technographics, scores each lead, and delivers the result in a format ready for your CRM or sales engagement tool.
What is the difference between the basic, standard, and comprehensive enrichment levels?
Basic returns company name, industry, size, and one or two contacts. Standard adds revenue, tech stack, and key buying signals. Comprehensive includes a full buying committee map, competitive intelligence, and personalized outreach recommendations per account.
Source text
generate-leads.md · Apache-2.0
---
name: generate-leads
description: Generate ICP-qualified leads with enrichment, intent signals, and export-ready outputs.
usage: /sales-prospecting:generate-leads --criteria "B2B SaaS" --count 50 --enrich comprehensive
---
# Generate Leads Command
Generate qualified leads based on ideal customer profile (ICP) criteria with enriched contact information and buying signals.
## Command Syntax
```bash
/sales-prospecting:generate-leads --criteria <criteria> --count <number> --enrich <level>
```
## Parameters
- `--criteria`: ICP definition (industry, size, tech stack, location, etc.)
- `--count`: Number of leads to generate (default: 20)
- `--enrich`: Enrichment level (basic|standard|comprehensive)
- `--intent`: Include intent signals (true|false, default: true)
- `--contacts`: Number of contacts per account (default: 3)
- `--format`: Output format (csv|json|salesforce|hubspot)
## Examples
### Basic Lead Generation
```bash
/sales-prospecting:generate-leads --criteria "SaaS companies, 50-500 employees, using Salesforce"
```
### Comprehensive Account Research
```bash
/sales-prospecting:generate-leads \
--criteria "Financial services, >$100M revenue, Northeast US" \
--count 50 \
--enrich comprehensive \
--contacts 5
```
### Intent-Based Targeting
```bash
/sales-prospecting:generate-leads \
--criteria "Recently funded B2B companies" \
--intent true \
--enrich standard
```
## Process Flow
1. **Define ICP Parameters**
- Parse criteria into searchable filters
- Validate parameter compatibility
- Set enrichment requirements
2. **Initial Discovery**
- Query data sources for matching companies
- Apply filtering and scoring logic
- Rank by fit score
3. **Contact Discovery**
- Identify decision makers and influencers
- Map buying committee structure
- Verify contact information
4. **Data Enrichment**
- Append firmographic data
- Add technographic insights
- Include intent signals and triggers
- Calculate lead scores
5. **Output Generation**
- Format according to specification
- Include research notes and insights
- Generate outreach recommendations
## Output Schema
### Standard Lead Record
```json
{
"company": {
"name": "Acme Corp",
"domain": "acme.com",
"industry": "Software",
"size": 250,
"revenue": "$50M",
"location": "San Francisco, CA",
"growth_rate": "45%"
},
"contacts": [
{
"name": "John Smith",
"title": "VP Sales",
"email": "[email protected]",
"linkedin": "linkedin.com/in/johnsmith",
"phone": "+1-415-555-0100",
"role": "Decision Maker"
}
],
"technology": {
"crm": "Salesforce",
"marketing": "HubSpot",
"relevant_tools": ["Slack", "Zoom", "Google Workspace"]
},
"signals": {
"intent_score": 82,
"buying_signals": [
"Researching competitors",
"Downloaded pricing guides",
"Attending industry events"
],
"trigger_events": [
"New VP Sales hired 2 months ago",
"Series B funding closed"
]
},
"recommendations": {
"approach": "Reference their recent funding and expansion plans",
"value_prop": "Focus on scalability and integration with Salesforce",
"timing": "High - active buying cycle"
}
}
```
## Enrichment Levels
### Basic
- Company name, website, industry
- Basic size and location data
- 1-2 contacts with titles
- Lead score
### Standard
- All basic fields plus:
- Revenue and growth data
- 3-4 contacts with full details
- Technology stack overview
- Key buying signals
### Comprehensive
- All standard fields plus:
- Full buying committee mapping
- Detailed technographic profile
- Intent data and signals
- Competitive intelligence
- Personalized outreach recommendations
- Account research brief
## Integration Options
### CRM Export
```bash
# Direct to Salesforce
/sales-prospecting:generate-leads --criteria "..." --format salesforce
# HubSpot compatible
/sales-prospecting:generate-leads --criteria "..." --format hubspot
```
### Sales Engagement
```bash
# Outreach.io format
/sales-prospecting:generate-leads --criteria "..." --format outreach
# SalesLoft format
/sales-prospecting:generate-leads --criteria "..." --format salesloft
```
## Advanced Options
### Exclusion Lists
```bash
--exclude-companies "competitor1.com,competitor2.com"
--exclude-existing true # Skip current CRM records
```
### Scoring Customization
```bash
--scoring-model "enterprise" # Use enterprise scoring model
--minimum-score 70 # Only include leads scoring 70+
```
### Geographic Targeting
```bash
--location "San Francisco Bay Area"
--timezone "PST"
--language "English"
```
## Best Practices
1. **Start Narrow**: Begin with specific criteria and expand based on results
2. **Validate ICP**: Test initial results against known good customers
3. **Iterate Quickly**: Refine criteria based on response rates
4. **Track Performance**: Monitor which criteria generate best leads
5. **Regular Refresh**: Re-run monthly to catch new matches
## Error Handling
- **No matches found**: Broaden criteria or adjust filters
- **Too many results**: Add more specific criteria
- **Missing data**: Adjust enrichment level or data sources
- **Rate limits**: Batch processing for large requests
---
*Execution model: claude-haiku-4-5 for data processing, claude-sonnet-4 for insights generation*
© gtmagents · Apache-2.0 · gtmagents/gtm-agents