LinkedIn SDR Plugin for Claude Code

Autonomous LinkedIn SDR workflows powered by Claude Code. Comment on topic-related posts, send personalized connection requests, like posts from your network, and build relationships — all through natural conversation.

This is a community template. Fork it, customize it for your use case, and contribute back.

What It Does

Capability	Description
Topic Commenting	Scans LinkedIn for posts matching your topics and drafts genuine, contextual comments
Smart Connections	Researches prospects and crafts personalized connection request notes
Feed Engagement	Strategically likes/reacts to posts from connections and target accounts
Prospect Research	Deep-dives into a prospect's profile, posts, and activity for personalization
Warmup Sequences	Multi-touch engagement sequences to warm up cold prospects over 1-2 weeks
Daily Workflow	One command (/daily-engage) runs the full scan → draft → approve → execute cycle

How It Works

The plugin uses LinkedIn's internal Voyager API (the same API the LinkedIn web app uses) via session cookies from your browser. All actions go through a human-in-the-loop approval flow — nothing is posted without your explicit approval.

You: "Find posts about product-led growth and draft some comments"
Claude: [scans LinkedIn, finds 5 relevant posts, drafts comments]
Claude: "Here are 5 comment drafts. Approve all, or tell me which to skip."
You: "Post 1, 3, and 5 look good. Skip 2 and 4."
Claude: [posts approved comments with random delays between them]

Quick Start

1. Install the plugin

claude plugin add github:YOUR_ORG/linkedin-sdr-plugin

Or clone and install locally:

git clone https://github.com/YOUR_ORG/linkedin-sdr-plugin.git
cd linkedin-sdr-plugin
claude plugin add ./linkedin-sdr

2. Configure your LinkedIn cookies

Run the setup skill:

/configure-auth

Or manually edit linkedin-sdr/settings/settings.json:

1. Log into LinkedIn in your browser
2. Open DevTools (F12) → Application → Cookies → linkedin.com
3. Copy li_at and JSESSIONID cookie values
4. Set the CSRF token (JSESSIONID without ajax: prefix and quotes)
5. Copy your User-Agent from the browser console: navigator.userAgent

3. Configure your targeting

Edit settings/settings.json:

{
  "targeting": {
    "topics": ["product-led growth", "customer onboarding", "B2B SaaS"],
    "hashtags": ["plg", "customersuccess", "saas"],
    "target_titles": ["VP Product", "Head of CS", "Director of Growth"],
    "target_companies": ["Acme Corp", "Beta Inc"],
    "exclude_companies": ["My Company", "Competitor Inc"],
    "industries": ["Software", "SaaS"]
  },
  "persona": {
    "name": "Your Name",
    "title": "Your Title",
    "company": "Your Company",
    "value_prop": "I help B2B SaaS teams improve user onboarding",
    "tone": "professional-friendly",
    "comment_style": "insightful"
  }
}

4. Run your first daily engagement

/daily-engage

Architecture

Agents

Agent	Type	Description
@feed-scanner	Leaf	Scans feed and topic searches for engagement-worthy posts
@topic-commenter	Orchestrator	Finds topic posts + drafts contextual comments
@connection-sender	Orchestrator	Researches prospects + drafts connection notes
@feed-engager	Leaf	Plans likes/reactions for connection posts
@prospect-researcher	Leaf	Deep-dives into a prospect's profile and activity
@comment-drafter	Leaf	Drafts comment options for a specific post
@engagement-tracker	Leaf	Reports on daily/weekly engagement stats
@warmup-engager	Orchestrator	Multi-touch warmup sequences over days

Commands

Command	Description
/daily-engage	Full daily SDR workflow — scan, draft, approve, execute
/engagement-report	View engagement stats and insights

Skills

Skill	Description
configure-auth	Interactive setup for LinkedIn cookies

Safety & Rate Limiting

This plugin is designed to keep your LinkedIn account safe:

• Daily limits — configurable caps on comments (15), likes (50), connections (20)
• Random delays — 30-120 seconds between actions (configurable)
• Active hours — only runs during configured business hours
• Per-person limits — max 3 interactions per person per week
• Human-in-the-loop — every action requires explicit approval before execution
• Engagement logging — full audit trail of every action taken

Default Daily Limits

Action	Default Limit	Max Safe Limit
Comments	15/day	30/day
Likes	50/day	100/day
Connection requests	20/day	40/day
Profile views	30/day	50/day