Vijay-Duke/mcp-atlassian

MCP Atlassian Server

A Model Context Protocol (MCP) server for integrating with Atlassian products (Confluence and Jira). This server provides tools for AI assistants to interact with Atlassian Cloud APIs, enabling document management, search, and export capabilities.

Features

Confluence Integration

• Read & Search: Access pages, spaces, and content
• Content Management: Create, update pages and comments
• Page Hierarchy: Navigate parent/child page relationships
• Export: Export pages as HTML or Markdown with embedded images
• Attachments: List, download, and upload attachments
• Labels: Manage page labels
• Users: Find and query users, track personal activity
• Personal Dashboard: View recent pages and mentions

Jira Integration

• Issues: Read and search issues, get personal tasks
• Projects: List and explore projects
• Boards & Sprints: List boards, view sprints, track active work
• Comments: Add comments to issues
• Issue Creation: Create new issues with custom fields
• User Management: Get current user details
• Personal Dashboard: View your open issues and sprint tasks

Installation

Option 1: Clone and Build (Recommended)

# Clone the repository
git clone https://github.com/Vijay-Duke/mcp-atlassian.git
cd mcp-atlassian

# Install dependencies
npm install

# Build TypeScript
npm run build

Option 2: Install from GitHub

# Install directly from GitHub
npm install -g github:Vijay-Duke/mcp-atlassian

# Or install in your project
npm install github:Vijay-Duke/mcp-atlassian

Option 3: NPM Registry

# Install globally
npm install -g mcp-atlassian

# Or install locally
npm install mcp-atlassian

Configuration

Environment Variables

Create a .env file in the root directory:

ATLASSIAN_BASE_URL=https://yourdomain.atlassian.net
ATLASSIAN_EMAIL=your-email@example.com
ATLASSIAN_API_TOKEN=your-api-token

Getting API Token

1. Log in to Atlassian Account Settings
2. Click "Create API token"
3. Give it a label and copy the token
4. Use this token in your .env file

MCP Settings Configuration

Add to your Claude Desktop config file:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

Option 1: After npm install -g

{
  "mcpServers": {
    "mcp-atlassian": {
      "command": "npx",
      "args": ["mcp-atlassian"],
      "env": {
        "ATLASSIAN_BASE_URL": "https://yourdomain.atlassian.net",
        "ATLASSIAN_EMAIL": "your-email@example.com",
        "ATLASSIAN_API_TOKEN": "YOUR_API_TOKEN"
      }
    }
  }
}

Option 2: From Local Clone

{
  "mcpServers": {
    "mcp-atlassian": {
      "command": "node",
      "args": ["/path/to/your/mcp-atlassian/dist/index.js"],
      "env": {
        "ATLASSIAN_BASE_URL": "https://yourdomain.atlassian.net",
        "ATLASSIAN_EMAIL": "your-email@example.com",
        "ATLASSIAN_API_TOKEN": "YOUR_API_TOKEN"
      }
    }
  }
}

Example with typical paths:

{
  "mcpServers": {
    "mcp-atlassian": {
      "command": "node",
      "args": ["~/projects/mcp-atlassian/dist/index.js"],
      "env": {
        "ATLASSIAN_BASE_URL": "https://yourdomain.atlassian.net",
        "ATLASSIAN_EMAIL": "your.email@company.com",
        "ATLASSIAN_API_TOKEN": "YOUR_API_TOKEN"
      }
    }
  }
}

Option 3: Direct from GitHub using uvx (Coming Soon)

You can run the server directly from GitHub without cloning:

{
  "mcpServers": {
    "mcp-atlassian": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/Vijay-Duke/mcp-atlassian.git", "mcp-atlassian"],
      "env": {
        "ATLASSIAN_BASE_URL": "https://yourdomain.atlassian.net",
        "ATLASSIAN_EMAIL": "your-email@example.com",
        "ATLASSIAN_API_TOKEN": "YOUR_API_TOKEN"
      }
    }
  }
}

Note: The uvx method requires the package to be properly configured for Python packaging. This is planned for a future release.

Available Tools

Confluence Tools

Jira Tools

Usage Examples

Export Confluence Page

// Export as HTML (raw content with embedded images)
{
  "tool": "export_confluence_page",
  "arguments": {
    "pageId": "123456789",
    "format": "html"
  }
}

// Export as Markdown with metadata
{
  "tool": "export_confluence_page",
  "arguments": {
    "pageId": "123456789",
    "format": "markdown"
  }
}

Search Confluence

{
  "tool": "search_confluence_pages",
  "arguments": {
    "cql": "space=DEV AND text~'architecture'",
    "limit": 10
  }
}

Create Jira Issue

{
  "tool": "create_jira_issue",
  "arguments": {
    "projectKey": "PROJ",
    "issueType": "Task",
    "summary": "Implement new feature",
    "description": "Detailed description here",
    "priority": "Medium"
  }
}

Get Your Sprint Tasks

// Get your tasks in the current sprint
{
  "tool": "get_my_tasks_in_current_sprint",
  "arguments": {
    "projectKey": "PROJ"
  }
}

// Get all your open issues
{
  "tool": "get_my_open_issues",
  "arguments": {
    "projectKeys": ["PROJ1", "PROJ2"],
    "maxResults": 50
  }
}

Work with Boards and Sprints

// List boards for a project
{
  "tool": "list_jira_boards",
  "arguments": {
    "projectKeyOrId": "PROJ",
    "type": "scrum"
  }
}

// Get active sprints for a board
{
  "tool": "list_jira_sprints",
  "arguments": {
    "boardId": 123,
    "state": "active"
  }
}

User-Specific Jira Operations

// Get user details
{
  "tool": "get_jira_user",
  "arguments": {
    "username": "john.doe"
  }
}

// Search issues by user involvement
{
  "tool": "search_jira_issues_by_user",
  "arguments": {
    "username": "john.doe",
    "searchType": "assignee",
    "status": "In Progress",
    "maxResults": 20
  }
}

// Get user's work logs
{
  "tool": "get_user_jira_worklog",
  "arguments": {
    "username": "john.doe",
    "startDate": "2024-01-01",
    "endDate": "2024-01-31",
    "projectKeys": ["PROJ1", "PROJ2"]
  }
}

// Track user activity
{
  "tool": "get_user_jira_activity",
  "arguments": {
    "username": "john.doe",
    "activityType": "all",
    "days": 7
  }
}

Content Format Support

Markdown ↔ Confluence Storage Format

The server automatically converts between Markdown and Confluence's storage format:

• Write content in Markdown when creating/updating pages
• Read pages in either storage format or converted to Markdown
• Preserves formatting, links, and structure

Export Formats

• HTML: Raw Confluence HTML with all images embedded as base64 data URIs
• Markdown: Clean Markdown with YAML frontmatter, includes metadata and embedded images

Development

# Run in development mode with console logging
npm run dev

# Build for production
npm run build

# Run built server in MCP mode (no console output)
npm start

# Run built server in development mode (with console output)
npm run start:dev

# Run tests
npm test

Logging Behavior

The server defaults to MCP mode where console output is disabled to prevent interference with the JSON-RPC protocol. Console logging is only enabled when MCP_SERVER_MODE=false is set, which is done automatically by the dev and start:dev scripts.

Project Structure

mcp-atlassian/
├── src/
│   ├── index.ts                 # Main server entry point
│   ├── types/                   # TypeScript type definitions
│   ├── confluence/
│   │   ├── handlers.ts          # Confluence API handlers
│   │   └── tools.ts             # Tool definitions
│   ├── jira/
│   │   ├── handlers.ts          # Jira API handlers
│   │   └── tools.ts             # Tool definitions
│   └── utils/
│       ├── http-client.ts       # Axios HTTP client setup
│       ├── content-converter.ts # Markdown ↔ Storage conversion
│       └── export-converter.ts  # HTML/Markdown export utilities
├── dist/                        # Compiled JavaScript
├── .env                         # Environment variables (not in git)
├── package.json
└── tsconfig.json

Security Notes

• API tokens are stored in environment variables, never in code
• Uses Basic Authentication with API tokens (not passwords)
• All requests are made over HTTPS
• Supports Atlassian Cloud only (not Server/Data Center)

Limitations

• No delete operations implemented (by design for safety)
• Export to PDF requires browser conversion (HTML → Print → PDF)
• Some Confluence macros may not convert perfectly to Markdown
• Rate limits apply based on Atlassian Cloud API limits

CI/CD Pipeline

This project uses GitHub Actions for continuous integration and deployment.

Workflows

🔄 Continuous Integration (ci.yml)

• Triggers: Push to main/develop, Pull requests
• Jobs:
  • Test: Runs tests on Node.js 18.x, 20.x, and 22.x
  • Build: Compiles TypeScript and validates the build
  • Lint: Type checking and security audit
  • Validate Package: Ensures package size and structure

📦 Publish to npm (publish.yml)

• Triggers: GitHub releases, Manual dispatch
• Features:
  • Automatic version bumping
  • npm publishing with provenance
  • GitHub release creation
  • Changelog updates

🔒 Security Scanning (security.yml)

• Triggers: Push to main, PRs, Weekly schedule
• Scans:
  • npm audit for vulnerabilities
  • CodeQL analysis
  • OWASP dependency check
  • Snyk security scanning (optional)

🏷️ Release Management (release.yml)

• Triggers: Version tags, Manual dispatch
• Features:
  • Automatic changelog generation
  • GitHub release creation
  • Build artifacts attachment
  • Release notes formatting

✅ PR Validation (pr-validation.yml)

• Triggers: Pull request events
• Checks:
  • Semantic PR title validation
  • PR size labeling
  • Auto-labeling based on files changed

🤖 Automated Dependency Updates

Dependabot is configured to:

• Check for npm dependency updates weekly
• Check for GitHub Actions updates weekly
• Group non-major updates together
• Create PRs with proper labels

Setting Up CI/CD

Required GitHub Secrets

1. NPM_TOKEN: npm authentication token for publishing

   • Generate at: https://www.npmjs.com/settings/YOUR_USERNAME/tokens
   • Required scopes: publish

2. SNYK_TOKEN (Optional): For Snyk security scanning

   • Get from: https://app.snyk.io/account

Branch Protection

Recommended branch protection rules for main:

• Require PR reviews before merging
• Require status checks to pass (CI tests)
• Require branches to be up to date
• Include administrators in restrictions

Local Development

Before pushing changes:

# Run tests locally
npm test

# Build the project
npm run build

# Check for security vulnerabilities
npm audit

# Type check
npx tsc --noEmit

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT License - see LICENSE file for details

Support

For issues and questions:

• Create an issue in the GitHub repository
• Check Atlassian API documentation for API-specific questions
• Review MCP documentation for protocol-related topics

Acknowledgments

Built with:

• Model Context Protocol (MCP)
• Atlassian REST APIs
• TypeScript, Node.js, Axios

Contributors

Thanks goes to these wonderful people (emoji key):

This project follows the all-contributors specification. Contributions of any kind welcome!