kaushik-linkedin-mcp
Enables Claude AI to interact with LinkedIn through browser automation, including profile reading, people and job search, company research, post publishing, and profile editing.
README
๐ kaushik-linkedin-mcp
A personal LinkedIn MCP server โ built from scratch. Not a fork. Not a clone. Written by Kaushik Muthukumaran.
This MCP server connects Claude AI directly to LinkedIn using browser automation. It gives Claude 8 tools to read profiles, search people and jobs, get company data, publish posts, edit your profile, and manage sessions โ all from a conversation with Claude.
๐ Table of Contents
- What You Can Do With This
- How to Clone and Run It
- Project Structure
- How It Works
- Tools Reference
- For People Cloning This Repo
- Security
- Troubleshooting
- Author
๐ What You Can Do With This
Once connected to Claude, you can have natural conversations that do real LinkedIn actions.
โ๏ธ Edit Your LinkedIn Profile By Just Telling Claude
Method 1 โ Tell Claude your details in chat:
"Change my headline to: Security Engineer | Zero Trust | CrowdStrike | 14K+ Endpoints" "Add a new job: Security Engineer at MasTec, promoted after 1 year" "Update my Sankara Nethralaya role to focus on vulnerability assessments and OSINT"
Claude opens Chrome, finds the right field, types your content, and saves it.
Method 2 โ Upload your resume and ask Claude to sync it:
"Here's my updated resume โ update my LinkedIn to match it"
Claude reads every section, compares it to your LinkedIn, and edits each section one by one.
Supports: .pdf, .docx, .txt
Method 3 โ Targeted corrections mid-conversation:
"Split my MasTec role into two entries at August 2025" "Add OSINT and CyberArk to my skills" "Move Sankara Nethralaya from Part-time to Full-time"
๐ค Profile Intelligence
"Pull Kaushik's full LinkedIn profile including experience and skills" "Show me their recent posts and engagement"
Reads: name, headline, location, connections, about, full experience, education, skills, posts, contact info.
๐ People Search
"Find Security Engineers at CrowdStrike in Pennsylvania" "Search for CISOs in the healthcare industry"
Returns: name, headline, location, username, profile URL. Max 100 results per call.
๐ผ Job Search & Research
"Find Senior Security Engineer jobs in Pennsylvania posted this week" "Show me remote Zero Trust architect roles โ mid-senior level only" "What's the salary range for this CrowdStrike job?"
Full LinkedIn job search: date posted, experience level, job type, work arrangement, Easy Apply.
๐ข Company Research
"What does CrowdStrike's LinkedIn page say about them?" "Show me MasTec's recent posts and engagement numbers" "What roles is Microsoft hiring for right now?"
Returns: about, industry, size, headquarters, recent posts with reactions, open roles.
๐ข Post Publishing
"Publish my vulnerability management post to LinkedIn" "Post this Zero Trust insight with these hashtags"
Navigates to feed, opens composer, injects text securely, clicks Post, confirms success.
๐ค Combined Power Workflows
"Search for Senior Security Engineer jobs at CrowdStrike, get the top 3 JDs, and tell me exactly how my resume matches each one"
"Find Security Engineers at my target companies and draft a personalised connection message for each one based on their background"
"Write and publish a post about today's Rapid7 CVE findings โ under 200 words with 5 hashtags"
๐ฅ๏ธ How to Clone and Run It
Step 1 โ Install Homebrew
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
echo 'eval "$(/opt/homebrew/bin/brew shellenv zsh)"' >> ~/.zprofile
eval "$(/opt/homebrew/bin/brew shellenv zsh)"
Step 2 โ Install Node.js and Git
brew install node git
Step 3 โ Install uv
curl -LsSf https://astral.sh/uv/install.sh | sh
source $HOME/.local/bin/env
โ ๏ธ Save the path from
which uvโ you'll need it for Claude Desktop config.
Step 4 โ Clone and Install
git clone https://github.com/Secure-tree/kaushik-linkedin-mcp
cd kaushik-linkedin-mcp
uv sync
uv run playwright install chromium
Step 5 โ Authenticate with LinkedIn
uv run kaushik-linkedin-mcp --login
A Chrome window opens. Log in to your LinkedIn. Session saves automatically.
Step 6 โ Verify Session
uv run kaushik-linkedin-mcp --status
# Status: Valid โ
Step 7 โ Connect to Claude Desktop
Open: ~/Library/Application Support/Claude/claude_desktop_config.json
Merge this in:
{
"mcpServers": {
"linkedin": {
"command": "/Users/YOUR_USERNAME/.local/bin/uv",
"args": [
"--directory",
"/Users/YOUR_USERNAME/kaushik-linkedin-mcp",
"run",
"kaushik-linkedin-mcp"
]
}
}
}
โ ๏ธ Use the full path to
uvfromwhich uvโ Claude Desktop does not inherit your terminal PATH.
Restart Claude:
pkill -f Claude && sleep 2 && open /Applications/Claude.app
๐๏ธ Project Structure
kaushik-linkedin-mcp/
โโโ pyproject.toml # Project config, CLI entry point, dependencies
โโโ .python-version # Pins Python 3.12
โโโ .env.example # Config template โ copy to .env
โโโ .gitignore # Protects session data and .env
โโโ LICENSE # MIT
โโโ src/
โโโ kaushik_linkedin_mcp/
โโโ __init__.py
โโโ server.py # FastMCP server โ all 8 tools
โโโ browser.py # Playwright session manager
โโโ tools/
โโโ profile.py # get_person_profile
โโโ jobs.py # search_jobs + get_job_details
โโโ posts.py # publish_linkedin_post
โโโ search.py # search_people + get_company_profile
โ๏ธ How It Works
You (in Claude chat)
โ
โ "Search for Security Engineer jobs in Pennsylvania"
โผ
Claude Desktop
โ stdio (local process)
โผ
kaushik-linkedin-mcp (Python / FastMCP)
โ Playwright browser automation
โผ
Chromium (headless, on your Mac)
โ HTTPS โ your authenticated session
โผ
LinkedIn.com โ returns structured JSON back to Claude
Note for cloners: Session saves to
~/.kaushik-linkedin-mcp/session/storage.jsonon YOUR machine. The author has zero access to your account.
๐งฐ Tools Reference
| Tool | Parameters |
|---|---|
get_person_profile |
linkedin_username, sections (main/experience/education/skills/posts/contact) |
search_linkedin_people |
keywords, location, max_results (1โ100) |
search_linkedin_jobs |
keywords, location, date_posted, experience_level, job_type, work_type, easy_apply, max_pages |
get_linkedin_job_details |
job_id |
get_linkedin_company |
company_name, sections (about/posts/jobs) |
publish_linkedin_post |
text |
check_session_status |
โ |
close_browser |
โ |
๐ฅ For People Cloning This Repo
When you clone and run this, you log into your own LinkedIn โ not the author's.
| Thing | Change Required |
|---|---|
--login step |
Log in with YOUR LinkedIn credentials |
| Claude Desktop config | Update path to YOUR clone location and uv path |
.env |
Optional โ defaults work for most setups |
pyproject.toml author |
Leave as-is โ this is the author's name |
Quick start:
git clone https://github.com/Secure-tree/kaushik-linkedin-mcp
cd kaushik-linkedin-mcp
uv sync
uv run playwright install chromium
uv run kaushik-linkedin-mcp --login
uv run kaushik-linkedin-mcp --status
๐ Security
| Check | Status |
|---|---|
| No hardcoded credentials | โ |
| No JS template literal injection | โ User values passed via CDP args |
| Input validation | โ max_results clamped 1โ100 |
| Session stored outside repo | โ
~/.kaushik-linkedin-mcp/ |
.env never committed |
โ In .gitignore |
| No shell/subprocess calls | โ |
| No 3rd party API calls | โ Only linkedin.com |
| Local stdio communication | โ No network exposure |
๐ง Troubleshooting
No LinkedIn tools in Claude
โ Use full uv path in config: which uv โ /Users/YOU/.local/bin/uv
Status: Expired/Invalid
โ uv run kaushik-linkedin-mcp --logout && uv run kaushik-linkedin-mcp --login
npm install fails
โ This is Python. Use uv sync not npm install.
Config has two {} blocks
โ Merge into one root object: { "preferences": {...}, "mcpServers": {...} }
โก Quick Reference
# Setup
git clone https://github.com/Secure-tree/kaushik-linkedin-mcp
cd kaushik-linkedin-mcp && uv sync
uv run playwright install chromium
uv run kaushik-linkedin-mcp --login
uv run kaushik-linkedin-mcp --status
# Restart Claude after config changes
pkill -f Claude && sleep 2 && open /Applications/Claude.app
# Re-authenticate
uv run kaushik-linkedin-mcp --logout && uv run kaushik-linkedin-mcp --login
๐ค Author
Kaushik Muthukumaran Security Engineer | AI + Cybersecurity Researcher
- ๐ LinkedIn
- ๐ง kaushikrmk01@gmail.com
- ๐ Ms in Cybersecurity
- ๐ CompTIA Network+ Certified
- ๐ฌ Research: AI-Driven Threat Detection ยท Zero Trust ยท ML Risk Scoring and Remediation
๐ License
MIT License โ see LICENSE for details.
Built with Claude AI (Anthropic) ยท FastMCP ยท Playwright ยท Python 3.12
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.