claude-mcp-setup

Claude MCP Servers

A collection of Model Context Protocol (MCP) servers that extend Claude with real-world capabilities like GitHub, Jira, file operations, and web access.

What is MCP?

Model Context Protocol (MCP) is a standard for connecting AI assistants to external tools and data sources. Think of it as a universal adapter that lets Claude interact with:

• APIs (GitHub, Jira, web services)
• File systems (read, write, search files)
• Databases (PostgreSQL, Redis)
• Command line (execute bash commands safely)

Why build with MCP? You'll learn:

• How AI assistants integrate with real systems
• Secure API design patterns
• Async programming with Python
• Error handling and validation
• Tool composition and orchestration

Quick Setup

# 1. Clone and setup
git clone https://github.com/souravs72/claude-mcp-setup.git
cd claude-mcp-setup

# 2. Install dependencies
make setup

# 3. Configure credentials
cp config/mcp_settings.json.template config/mcp_settings.json
# Edit config/mcp_settings.json with your API keys

# 4. Start servers
make start

Available Servers

Server	Purpose	Key Features
GitHub	Repository management	Create repos, manage PRs, handle issues
Jira	Project management	Create issues, manage sprints, story points
File	File operations	Read/write files, search, directory listing
Bash	Command execution	Safe command running with validation
Internet	Web access	Google search, fetch web content
Memory	Caching	Redis-based data storage
Frappe	ERP integration	Connect to Frappe/ERPNext systems

Configuration

Edit config/mcp_settings.json with your credentials:

{
  "mcpServers": {
    "github-server": {
      "env": {
        "GITHUB_PERSONAL_ACCESS_TOKEN": "your_token_here"
      }
    },
    "jira-server": {
      "env": {
        "JIRA_BASE_URL": "https://yourcompany.atlassian.net",
        "JIRA_EMAIL": "your@email.com",
        "JIRA_API_TOKEN": "your_token_here",
        "JIRA_PROJECT_KEY": "PROJ"
      }
    }
  }
}

Usage

Once configured, Claude can:

# GitHub operations
"Create a new repository called 'my-project'"
"List all open pull requests"
"Create an issue with title 'Bug fix needed'"

# Jira operations
"Create a new task with 5 story points"
"List all issues in the current sprint"
"Update issue PROJ-123 status to 'In Progress'"

# File operations
"Read the contents of README.md"
"Search for all Python files in the project"
"Create a new file called 'config.py'"

# Web operations
"Search Google for 'Python async programming'"
"Fetch the content from https://example.com"

Development

# Check server status
make status

# View logs
make logs

# Run tests
make test

# Stop all servers
make stop

Key Features

• 🔒 Security First: Command validation, path restrictions, API key protection
• 🤖 AI-Friendly: Structured error messages, helpful suggestions
• ⚡ Fast: Async operations, efficient caching
• 🛠️ Developer-Friendly: Clear APIs, comprehensive logging
• 🔧 Extensible: Easy to add new servers and capabilities

Requirements

• Python 3.10+
• Redis (for caching)
• PostgreSQL (for persistent storage)
• API keys for external services

License

MIT License - see LICENSE for details.

---

Learning MCP? This project demonstrates real-world MCP server implementation with production-ready patterns for security, error handling, and AI integration.