Jenkins MCP Server

A Python-based Jenkins MCP server using the Model Context Protocol Python SDK. This server integrates with Jenkins CI/CD systems to provide AI-powered insights, build management, and debugging capabilities.

Note:This server follows theModel Context Protocol (MCP), enabling AI assistants to interact with Jenkins systems directly.

# Install from PyPI pip install jenkins-mcp-server==0.1.6 # Or install with uv uv pip install jenkins-mcp-server==0.1.6 # Run the server jenkins-mcp-server --verbose

# Clone the repository git clone https://github.com/yourusername/jenkins-mcp-server.git cd jenkins-mcp-server # Create a virtual environment and install dependencies uv venv uv pip install -e . # Run the server python -m jenkins_mcp_server

{ "servers": { "jenkins-mcp-server": { "type": "stdio", "command": "jenkins-mcp-server", "args": ["--verbose"], "env": { "JENKINS_URL": "http://your-jenkins-server:8080", "JENKINS_USERNAME": "your-username", "JENKINS_TOKEN": "your-api-token" // Or use JENKINS_PASSWORD instead of JENKINS_TOKEN if using password authentication } } } }

- Recommended: Use API token authentication by settingJENKINS_TOKEN
- Alternatively: Use password authentication by settingJENKINS_PASSWORD

Connect any AI assistant that supports MCP (like GitHub Copilot) to your Jenkins environment

The server provides access to Jenkins jobs as resources:

- Custom jenkins:// URI scheme for accessing individual jobs
- Each job resource contains details about the job and its builds in JSON format
- Job status is reflected in the resource description

The server provides prompts for Jenkins data analysis:
-

analyze-job-status: Creates analysis of all Jenkins jobs

- Optional "detail_level" argument to control analysis depth (brief/detailed)
- Analyzes job statuses, identifies potential issues, and suggests improvements

analyze-build-logs: Analyzes build logs for a specific job

- Required "job_name" argument to specify which job to analyze
- Optional "build_number" argument (defaults to latest build)
- Examines build logs to identify issues, errors, warnings, and suggests fixes

The server implements the following tools for Jenkins operations:
-

trigger-build: Triggers a Jenkins job build

- Required "job_name" argument to specify which job to build
- Optional "parameters" object containing job parameters
- Returns build queue information

stop-build: Stops a running Jenkins build

- Required "job_name" and "build_number" arguments
- Halts an in-progress build execution

get-job-details: Gets detailed information about a specific job

- Required "job_name" argument
- Returns comprehensive job information including recent builds

- Returns a list of all Jenkins jobs with their statuses

get-build-info: Gets information about a specific build

- Required "job_name" and "build_number" arguments
- Returns build status, duration, and other details

get-build-console: Gets console output from a build

- Required "job_name" and "build_number" arguments
- Returns the console log output from a specific build

get-queue-info: Gets information about the Jenkins build queue

- Returns information about pending builds in the queue

get-node-info: Gets information about a Jenkins node/agent

- Required "node_name" argument
- Returns node status and configuration details

list-nodes: Lists all Jenkins nodes/agents

- Returns a list of all Jenkins nodes/agents and their statuses

Option 1: VS Code Settings (Recommended)

Configure your Jenkins connection in VS Code settings:
- Open VS Code Settings (PressCmd+,on Mac orCtrl+,on Windows/Linux)
- Click on the "Open Settings (JSON)" button in the top right
- Add these settings:

{ "mcp.servers": { "jenkins": { "type": "stdio", "command": "uvx", "args": ["jenkins-mcp-server"] } }, "jenkins-mcp-server.jenkins": { "url": "http://your-jenkins-server:8080", "username": "your-username", "token": "" // Replace with your Jenkins API token } }

{ "servers": { "jenkins-mcp-server": { "type": "stdio", "command": "jenkins-mcp-server", "args": [ "--verbose" ] } } }

{ "jenkins-mcp-server.jenkins": { "url": "http://your-jenkins-server:8080", "username": "your-username", "token": "" // Replace with your Jenkins API token } }

- Registers the MCP server in VS Code
- Stores your Jenkins credentials securely in VS Code settings
- Usesuvxto run the server automatically when needed

Alternatively, configure your Jenkins connection by setting environment variables:
-

Copy the.env.examplefile to create a.envfile:

Edit the.envfile with your Jenkins details:

JENKINS_URL=http://your-jenkins-server:8080 JENKINS_USERNAME=your-username JENKINS_PASSWORD=your-password # OR use an API token instead of password (recommended) JENKINS_TOKEN=your-api-token

Security Note:Using VS Code settings is more secure as they are stored encrypted. Environment variables in a.envfile are stored in plain text.

Once configured, AI assistants that support MCP can now interact with your Jenkins server through natural language. Here are some examples of what you can do:
- Open GitHub Copilot Chat in VS Code
- Type prompts like:

- "List all my Jenkins jobs"
- "What's the status of my 'deployment' job?"
- "Show me the build logs for the failed build in 'test-project'"
- "Trigger a new build for 'deploy-api'"

You can also run the server directly from the command line:

# Run the MCP server uvx jenkins-mcp-server # In another terminal, use curl to test it: curl -X POST http://localhost:8080/mcp/v1/listResources -H "Content-Type: application/json" -d '{}'

Theuvxcommand makes it easy to use the MCP server in command-line environments without VS Code:

# Install UVX if you don't have it yet pip install uv # Install the Jenkins MCP server uvx install jenkins-mcp-server 

You can manually send JSON-RPC requests:

If you're developing this MCP server:

- Clone this repository

- Install dependencies:

- Install the GitHub Copilot Chat extension

- Enable MCP in Copilot settings

- Start a new chat with Copilot and interact with your Jenkins server!

On MacOS:~/Library/Application\ Support/Claude/claude_desktop_config.jsonOn Windows:%APPDATA%/Claude/claude_desktop_config.json

To prepare the package for distribution:

This will create source and wheel distributions in thedist/directory.
Note: You'll need to set PyPI credentials via environment variables or command flags:
- Token:--tokenorUV_PUBLISH_TOKEN

- Or username/password:--username/UV_PUBLISH_USERNAMEand--password/UV_PUBLISH_PASSWORD
Since MCP servers run over stdio, debugging can be challenging. For the best debugging experience, we strongly recommend using theMCP Inspector.

You can launch the MCP Inspector vianpmwith this command:

Debug your remote Node.js and Next.js applications directly from your AI IDE like Cursor.