Outreach Engine — Testing skill for Claude Code

outreach-engine

Built and run for an AI ad startup's outbound. Published as deployed, minus the real ICP profiles, the competitor skip list and the lead data. Copy `config/icp_profiles.example.json` to `config/icp_profiles.json` and describe your own segments; everything else reads from `.env`.

Two pipelines in one repo:

1. Sales Prospecting — Find potential customers (ICP-based), enrich via Apollo, score, push to Instantly
2. Link Building — Find SEO link opportunities (listicles, mentions, competitor backlinks), score by DR/relevance, personalize via Claude, push to Instantly

Both share the same Instantly → n8n → reply triage infrastructure.

---

System Architecture

Phase 1: Claude Code           Phase 2: n8n               Phase 3: Instantly
┌─────────────────────┐       ┌───────────────────┐       ┌──────────────────┐
│ Prospecting Engine  │──CSV─▶│ CSV → Instantly   │──API─▶│ Drip sequences   │
│                     │       │ push pipeline     │       │ Inbox rotation   │
│ - Apollo enrichment │       │ (every 15 min)    │       │ Timezone sched.  │
│ - Crunchbase signals│       └───────────────────┘       │                  │
│ - Tech stack detect │                                   │ You manage:      │
│ - ICP scoring       │       Phase 4: n8n + Instantly    │ - Copy           │
│ - QA check          │       ┌───────────────────┐       │ - Subject lines  │
│ - CSV write         │◀──────│ Reply webhooks     │       │ - Sequences      │
└─────────────────────┘       │ LLM triage        │       └──────────────────┘
                               │ leads_master.csv  │
                               └───────────────────┘

---

Quick Start

1. Setup

cd outreach-engine

# Install dependencies
pip install -r requirements.txt

# Copy and fill in your API keys
cp .env.example .env
# Edit .env with your actual keys

2. Run Phase 1 (Prospecting)

# Process the seeded AI SaaS list
python scripts/run_prospecting.