lead-gen-ai

name	lead-gen-ai
description	Automated lead extraction system — Vibe Prospecting MCP + Apollo MCP + Excel export + outreach sequences. Also orchestrates hmz-daily-leads, hmz-bdm-morning-sweep, hmz-bdm-evening-sweep, hmz-indeed-mcp-sweep scheduled pipelines.
allowed-tools	Read, Bash, Write, Agent

Lead Gen AI — Automated Lead Extraction System

Triggers: lead generation, find leads, extract leads, lead gen, vibe prospecting, find contacts, scrape leads, dental clinic leads, business directory, find emails, find phone numbers, contact scraping, automated leads, apollo leads, clay leads, hmz leads, daily leads, elite leads, lead sweep, bdm sweep

---

THE SYSTEM — One prompt → full contact dataset

Name, phone, email, website, LinkedIn, Instagram, Facebook, Google rating Runs 24/7. Works for ANY niche, ANY location worldwide.

PLUS: Autonomous daily sweeps running on schedule for HMZ elite B2B leads.

---

MASTER AGENT WORKFLOW — 5 ACTORS (General Lead Gen)

ACTOR 1 — Brief Generator (Kimi 8K)

Input: "find 20 dental clinics in Dubai"
Output: structured prompt → business type + location + data fields + output format

ACTOR 2 — Vibe Prospecting Extractor (MCP)

mcp__Vibe_Prospecting__fetch-entities          → search by type + location
mcp__Vibe_Prospecting__enrich-prospects        → email, phone, social profiles
mcp__Vibe_Prospecting__fetch-businesses-events → buying signals (funding, hiring)
mcp__Vibe_Prospecting__match-business          → exact company match
mcp__Vibe_Prospecting__export-to-csv           → download dataset

ACTOR 3 — Apollo Enricher (MCP)

apollo_mixed_companies_search   → find companies by industry + filters
apollo_people_match             → find decision makers by name + company
apollo_contacts_search          → filter contacts by title, industry, location
apollo_organizations_enrich     → full company profile + tech stack
apollo_people_bulk_match        → batch enrich 100+ contacts

ACTOR 4 — Data Formatter + Excel Export (Python/openpyxl)

import openpyxl
from pathlib import Path

def export_leads_to_excel(leads: list[dict], filename: str = "leads"):
    wb = openpyxl.Workbook()
    ws = wb.active
    ws.title = "Leads"
    headers = ["Business Name", "Address", "Phone", "Email", "Website",
               "LinkedIn", "Instagram", "Facebook", "Google Rating", "Reviews", "Source"]
    for col, h in enumerate(headers, 1):
        ws.cell(1, col, h).font = openpyxl.styles.Font(bold=True)
    for row, lead in enumerate(leads, 2):
        for col, key in enumerate(["name","address","phone","email","website",
                                    "linkedin","instagram","facebook","rating","reviews","source"], 1):
            ws.cell(row, col, lead.get(key, ""))
    for col in ws.columns:
        ws.column_dimensions[col[].column_letter].width = 
    out = Path.home() /  / 
    wb.save(out)
     (out)

ACTOR 5 — Outreach Sequence Generator (Kimi K2.5)

Input: lead list
Output:
  - Personalized email subject + body per lead
  - LinkedIn connection request message
  - Cold DM for Instagram/Facebook
  - Mailchimp/GetResponse import-ready CSV

---

HMZ DAILY SWEEP PIPELINE — Autonomous B2B Lead System

Four scheduled tasks that run automatically, all using the same MCP stack:

SWEEP 1 — Elite Leads (7 AM PKT daily)

Task: hmz-daily-leads File: /Users/mc/.claude/scheduled-tasks/hmz-daily-leads/SKILL.md

Sources (run in parallel):

Source	Tool	What it finds
Apollo MCP	apollo_mixed_companies_search	B2B companies by industry + size
Vibe MCP	fetch-businesses-events	Funding/hiring signals = buying signals
Apify MCP	call-actor	Meta Ad Library scrape, LinkedIn signals, Shopify stores
Meta Ad Library	WebSearch	Confirmed active ad spenders
Reddit/Twitter	WebSearch	Companies expressing PPC pain right now
Job postings	WebSearch	Companies posting PPC roles = have budget

Gate system (ALL must pass):

• GATE 1: Real + active business
• GATE 2: Spends on ads (or clearly needs to)
• GATE 3: Has a problem Zulqarnain solves
• GATE 4: Can pay ($5k+ ad spend/mo signal)
• GATE 5: Can hire globally/remote

Scoring threshold: ≥80 only (0-100 scale)

Output:

• /Users/mc/Downloads/Leads/HMZ-Leads-Daily.xlsx — appended (never wiped)
• /Users/mc/Downloads/HMZ-Elite-Leads-[DATE].html — gold/green color-coded cards
• Gmail draft → [e-mail]

SWEEP 2 — Morning Job Sweep (9 AM PKT daily)

Task: hmz-bdm-morning-sweep File: /Users/mc/.claude/scheduled-tasks/hmz-bdm-morning-sweep/SKILL.md

10 platforms in parallel: LinkedIn · WeWorkRemotely · Remotive · Remote.co · Wellfound · Contra · Pangian · Himalayas · X/Twitter · Jobspresso

Filters: Meta/Google Ads scope · $15+/hr · GEO whitelist (USA/UK/CA/AU/AE/EU) · GEO blacklist (IN/PK/BD/PH/IL) · <48h · Remote + PK-eligible Threshold: ≥65 score

Output: /Users/mc/Downloads/HMZ-BDM-Morning-[DATE].html + Gmail draft

SWEEP 3 — Evening Job Sweep (9 PM PKT daily)

Task: hmz-bdm-evening-sweep File: /Users/mc/.claude/scheduled-tasks/hmz-bdm-evening-sweep/SKILL.md

Same 10 platforms + Reddit value comments (3 max, strict cap)

Output: /Users/mc/Downloads/HMZ-BDM-Evening-[DATE].html + Reddit drafts + Gmail

SWEEP 4 — Indeed MCP Sweep (on-demand / scheduled)

Task: hmz-indeed-mcp-sweep File: /Users/mc/.claude/scheduled-tasks/hmz-indeed-mcp-sweep/SKILL.md

8 Indeed MCP queries via search_jobs + get_job_details — structured data only, no WebSearch

Output: /Users/mc/Downloads/HMZ-Indeed-MCP-[DATE]-[TIME].html + Gmail draft

---

State Tracking

All sweeps must call this on completion:

python3 /Users/mc/.claude/bin/hmz-bdm-state-update [sweep-name]
# sweep-name: hmz-daily-leads | hmz-bdm-morning-sweep | hmz-bdm-evening-sweep | hmz-indeed-mcp-sweep

Catch-up script (run if a sweep was missed):

python3 /Users/mc/.claude/bin/hmz-bdm-catchup

---

COMBINED MCP STACK (all wired, all available)

Vibe Prospecting MCP  → business discovery, enrichment, buying signals
Apollo MCP            → B2B contacts, company enrichment, email sequences
Apify MCP             → web scraping (Meta Ad Library, LinkedIn, Shopify)
Gmail MCP             → create_draft, send reports
Indeed MCP            → structured job search via search_jobs + get_job_details
WebSearch             → real-time signals (Reddit, Twitter, job boards)

---

QUICK START PROMPTS

# General lead extraction
"Find top 20 dental clinics in Dubai. Get phone, email, Instagram, Google rating. Export Excel."

# Run daily elite leads sweep now
"Run hmz-daily-leads sweep — find 10 elite B2B leads for Zulqarnain, score ≥80 only"

# Morning job sweep now
"Run HMZ morning BDM sweep — all 10 platforms, strict filters, score ≥65"

# Enrich with decision makers
"For each business, find the owner/marketing manager name + email using Apollo"

# Generate outreach
"Write personalized cold email + LinkedIn DM for each lead. Export CSV for Mailchimp."

---

OUTPUT FILES

File	Path	Format
General leads	~/Downloads/leads.xlsx	11-column Excel
Elite B2B leads (daily append)	~/Downloads/Leads/HMZ-Leads-Daily.xlsx	Score, company, DM, pain point, entry angle
Morning job report	~/Downloads/HMZ-BDM-Morning-[DATE].html	Scored job cards
Evening job report	~/Downloads/HMZ-BDM-Evening-[DATE].html	Scored cards + Reddit drafts
Indeed MCP report	~/Downloads/HMZ-Indeed-MCP-[DATE].html	Structured job cards
Elite leads HTML	~/Downloads/HMZ-Elite-Leads-[DATE].html	Gold/green lead cards
Outreach CSV	~/Downloads/outreach.csv	Mailchimp import-ready

---

EMAIL AUTOMATION

After export → upload to:

1. Mailchimp — drag CSV → create campaign → auto-personalize subject
2. GetResponse — autoresponder sequence → drip campaign
3. Apollo sequences — apollo_emailer_campaigns_add_contact_ids via MCP

Template:

Subject: Increase your [BUSINESS] appointments with [SERVICE]

Hi [NAME],

I noticed [SPECIFIC OBSERVATION].
We help [NICHE] businesses in [CITY] get [RESULT] in [TIMEFRAME].
[PROOF — client result or free audit offer]

Would a quick 15-min call this week make sense?