employee-mcp-server
employee-mcp-server
A Node.js MCP (Model Context Protocol) server that connects Claude to a MySQL employees database, exposing tools to list, search, add, and analyze employee records.
Features
list_employees — returns all employees
search_employee — find employees by name (partial match)
employees_by_department — filter employees by department
add_employee — insert a new employee (name, department, salary)
highest_salary — returns the highest-paid employee
Tech Stack
Node.js (ES modules)
@modelcontextprotocol/sdk — MCP server implementation
mysql2 — MySQL driver
zod — input schema validation
dotenv — local environment variable loading
Project Structure
employee-mcp/
├── index.js # MCP server: defines and registers the 5 tools
├── db.js # MySQL connection pool
├── manifest.json # Desktop Extension (.mcpb) manifest
├── package.json # Dependencies
├── .env # Local DB credentials (not committed)
└── .gitignore
Setup
Install dependencies:
bash
npm install
Create a .env file in the project root:
DB_HOST=localhost
DB_USER=root
DB_PASSWORD=your_password
DB_NAME=company
Run the server directly (for testing):
bash
node index.js
Using with Claude Desktop
This project is packaged as a Desktop Extension (.mcpb) for one-click install:
bash
npm install -g @anthropic-ai/mcpb
npm install --production
mcpb validate manifest.json
mcpb pack .
Then in Claude Desktop: Settings → Extensions → Advanced settings → Extension Developer → Install Extension…, select the generated .mcpb file, and enter your MySQL credentials in the setup screen.
Example Prompts
Prompt Tool Called
"List all employees" list_employees
"Search for an employee named Mahesh" search_employee
"Show employees in the Engineering department" employees_by_department
"Add a new employee named Mahesh Khatal in Engineering with salary 120000" add_employee
"Who has the highest salary?" highest_salary
Notes
.env and node_modules/ are excluded via .gitignore and never committed.
Full build walkthrough and debugging notes available in the accompanying project documentation.
TDQS
Scored across 5 tools
Each tool has a clear purpose: list all, search by name, filter by department, add, and find highest salary. While search and department filtering overlap with listing, the descriptions clarify specific query intents, so ambiguity is minimal.
Tool names follow mixed conventions: list_employees, search_employee, and add_employee use verb_noun, but employees_by_department and highest_salary do not follow that pattern. The names are still readable and descriptive, but not fully consistent.
With 5 tools, the server is well-scoped for employee management. Each tool covers a distinct need without excessive fragmentation or bloat, fitting comfortably in the ideal range.
Provides read and create operations (list, search, department filter, add, highest salary) but lacks update and delete operations. This is a notable gap for a full employee management lifecycle.