AgentSkillsCN

OSINT Tradecraft

当用户请求“研究某个人”“查找某人的信息”“查询某人”“收集某人的情报”“展开调查”“寻找联系信息”“获取电子邮件”“丰富潜在客户”“识别决策者”“研究某家公司”“建立潜在客户名单”,或在销售线索挖掘与潜在客户开发中需要开源情报收集方面的指导时,应使用此技能。

SKILL.md
--- frontmatter
name: OSINT Tradecraft
description: This skill should be used when the user asks to "research a person", "find info on", "look up", "gather intelligence on", "investigate", "find contact info", "find email", "enrich leads", "identify decision makers", "research a company", "build prospect list", or needs guidance on open-source intelligence gathering for sales prospecting and lead generation.
version: 1.0.0

OSINT Tradecraft for Sales Intelligence

Comprehensive methodology for gathering personal and business intelligence using open-source techniques. This skill transforms raw names and company identifiers into actionable sales intelligence.

Core Principles

Data Source Hierarchy

Query sources in order of reliability and cost-effectiveness:

  1. Hunter.io (highest confidence for email)

    • Domain search for company emails
    • Email verification
    • Email finder with confidence scores
  2. Google Maps/Places API (business verification)

    • Confirm company exists and is operational
    • Get verified address, phone, hours
    • Reviews indicate company health
  3. SerpAPI (web search aggregation)

    • LinkedIn profiles (public data)
    • Company news and press releases
    • Social media presence
    • Publications and speaking engagements
  4. Direct web scraping (supplementary)

    • Company websites for team pages
    • Press releases for executive changes
    • Job postings for org structure clues

Verification Protocol

Never trust single-source data. Cross-reference using:

  • Email confidence: Hunter.io provides confidence scores (accept >80%)
  • Phone verification: Google Maps confirms business phone
  • Title validation: LinkedIn + company website agreement
  • Recency check: Prefer data updated within 6 months

Person Intelligence Workflow

Step 1: Initial Search

Start with the most specific identifiers available:

  • Full name + company
  • Full name + location
  • Email address (if known)
  • LinkedIn URL (if known)

Step 2: Email Discovery

Use Hunter.io's email finder:

code
Domain: company.com
First name: John
Last name: Smith
→ john.smith@company.com (95% confidence)

Verify discovered emails before storing.

Step 3: Professional Profile

Gather from SerpAPI LinkedIn search:

  • Current title and company
  • Previous roles (work history)
  • Education background
  • Skills and endorsements
  • Mutual connections

Step 4: Social Presence

Search for additional profiles:

  • Twitter/X (industry thought leadership)
  • GitHub (technical roles)
  • Medium/Substack (content creators)
  • Speaking engagements (conference sites)

Step 5: Decision-Maker Scoring

Calculate a 1-10 decision-maker score based on:

FactorPoints
C-level title+4
VP/Director title+3
Manager title+2
"Head of" or "Lead"+2
Budget keywords in title+1
5+ years at company+1
Previous purchasing roles+1

Company Intelligence Workflow

Step 1: Basic Verification

Use Google Maps/Places API to confirm:

  • Company exists at stated location
  • Current operating status
  • Verified contact information
  • Review sentiment analysis

Step 2: Company Profile

Gather from multiple sources:

  • Industry classification (SIC/NAICS)
  • Employee count range
  • Revenue estimate (if available)
  • Founded date
  • Headquarters location
  • Additional office locations

Step 3: Leadership Identification

Build executive roster:

  1. Search "[Company] leadership team"
  2. Search "[Company] executives"
  3. Check company website /about or /team pages
  4. Cross-reference with LinkedIn company page

Step 4: Employee Roster

For full org intel:

  1. Hunter.io domain search (returns all known emails)
  2. SerpAPI LinkedIn company employees search
  3. Categorize by department and seniority
  4. Identify reporting structures where possible

Step 5: Decision-Maker Mapping

Identify buyers with authority:

  • Find people with "purchasing", "procurement", "vendor" in title
  • Determine who manages the relevant department
  • Map C-level stakeholders
  • Document the typical buying committee structure

Step 6: Competitive Intelligence

Research market position:

  • Direct competitors (same product/service)
  • Recent news (funding, launches, pivots)
  • Job postings (indicate growth areas)
  • Tech stack (BuiltWith, Wappalyzer)

Output Formats

Person Intelligence Report

markdown
# Person Intelligence: [Full Name]

## Contact Information
- **Email**: verified@company.com (95% confidence)
- **Phone**: +1-555-123-4567 (business)
- **LinkedIn**: linkedin.com/in/username
- **Twitter**: @handle

## Professional Profile
- **Current Role**: VP of Engineering at Acme Corp
- **Tenure**: 3 years, 4 months
- **Location**: San Francisco, CA

## Work History
1. VP Engineering, Acme Corp (2022-present)
2. Director Engineering, Previous Co (2019-2022)
3. Senior Engineer, First Job (2015-2019)

## Education
- MS Computer Science, Stanford University
- BS Engineering, UC Berkeley

## Social Presence
- Active on Twitter (12k followers)
- GitHub contributor (open source projects)
- Conference speaker (DevCon 2024)

## Decision-Maker Score: 8/10
- VP title (+3)
- 3+ years tenure (+1)
- Technical budget authority implied (+2)
- Previously evaluated vendors (+2)

Company Intelligence Report

markdown
# Company Intelligence: [Company Name]

## Overview
- **Industry**: Enterprise Software (SIC: 7372)
- **Founded**: 2015
- **Employees**: 250-500
- **Revenue**: $50M-$100M (estimated)
- **Funding**: Series C ($45M)

## Location
- **HQ**: 123 Main St, San Francisco, CA 94102
- **Phone**: +1-555-000-0000
- **Website**: https://company.com

## Leadership
| Name | Title | Email | DM Score |
|------|-------|-------|----------|
| Jane CEO | CEO | jane@company.com | 10 |
| John CTO | CTO | john@company.com | 9 |
| Sarah VP Sales | VP Sales | sarah@company.com | 8 |

## Department Breakdown
- Engineering: 120 employees
- Sales: 45 employees
- Marketing: 25 employees
- Operations: 35 employees

## Decision Makers (Your Product Category)
1. **Primary**: Sarah VP Sales (budget holder)
2. **Technical**: John CTO (technical approval)
3. **Executive**: Jane CEO (final sign-off)

## Recent News
- [2024-01] Announced Series C funding
- [2024-02] Launched new product line
- [2024-03] Expanded to European market

## Competitors
- Competitor A (larger, enterprise focus)
- Competitor B (similar size, SMB focus)
- Competitor C (emerging, AI-native)

API Integration Patterns

Hunter.io

python
# Email finder
GET https://api.hunter.io/v2/email-finder
  ?domain=company.com
  &first_name=John
  &last_name=Smith
  &api_key=$HUNTER_API_KEY

# Domain search (all emails)
GET https://api.hunter.io/v2/domain-search
  ?domain=company.com
  &api_key=$HUNTER_API_KEY

Google Maps/Places

python
# Place search
GET https://maps.googleapis.com/maps/api/place/findplacefromtext/json
  ?input=Acme+Corp+San+Francisco
  &inputtype=textquery
  &fields=name,formatted_address,formatted_phone_number
  &key=$GOOGLE_MAPS_KEY

# Place details
GET https://maps.googleapis.com/maps/api/place/details/json
  ?place_id=ChIJ...
  &fields=name,formatted_address,formatted_phone_number,website,reviews
  &key=$GOOGLE_MAPS_KEY

SerpAPI

python
# LinkedIn profile search
GET https://serpapi.com/search
  ?engine=google
  &q=John+Smith+VP+Engineering+site:linkedin.com
  &api_key=$SERPAPI_KEY

# Company news search
GET https://serpapi.com/search
  ?engine=google
  &q="Acme+Corp"+news
  &tbs=qdr:m  # Past month
  &api_key=$SERPAPI_KEY

Data Quality Standards

Confidence Thresholds

Data TypeAcceptVerifyReject
Email>80%50-80%<50%
PhoneVerifiedGoogle MapsUnverified
TitleMulti-sourceSingle sourceOutdated
AddressGoogle MapsCompany siteUnverified

Freshness Requirements

  • Contact info: Verify if >6 months old
  • Job titles: May change frequently, cross-reference
  • Company data: Revenue/size estimates valid ~1 year
  • News: Focus on past 6 months for relevance

Additional Resources

Reference Files

For detailed techniques and patterns:

  • references/data-sources.md - Complete API documentation and rate limits
  • references/scraping-patterns.md - Web scraping techniques and selectors

Scripts

Utility scripts in $CLAUDE_PLUGIN_ROOT/scripts/:

  • format-report.py - Format intelligence into markdown/JSON/CSV
  • export-leads.py - Export batch results to various formats