MCP Quick Start — 5 Minutes#

Get DevOps-OS running as an MCP server in Claude Desktop in under 5 minutes.

What You’ll Do#

1. Clone the repo
2. Install dependencies
3. Add to Claude Desktop config
4. Generate your first pipeline with AI

Total time: ~4 minutes

Step 1: Clone & Install (1 minute)#

git clone https://github.com/cloudengine-labs/devops_os.git
cd devops_os

# Create virtual environment
python3 -m venv .venv
source .venv/bin/activate  # macOS/Linux
# .venv\Scripts\activate   # Windows

# Install MCP dependencies
pip install -r mcp_server/requirements.txt

Step 2: Configure Claude Desktop (2 minutes)#

Find Your DevOps-OS Path#

# In the devops_os directory, get the full path
pwd
# Output: /Users/alice/projects/devops_os (macOS/Linux)
# Or: C:\Users\alice\projects\devops_os (Windows)

Open Claude Config File#

macOS:

open ~/Library/Application\ Support/Claude/claude_desktop_config.json

Windows: Open %APPDATA%\Claude\claude_desktop_config.json in your text editor

Linux:

nano ~/.config/Claude/claude_desktop_config.json

Add DevOps-OS to Config#

Insert this block (if mcpServers already exists, just add the devops-os entry):

{
  "mcpServers": {
    "devops-os": {
      "command": "python",
      "args": ["-m", "mcp_server.server"],
      "cwd": "/Users/alice/projects/devops_os"
    }
  }
}

Replace /Users/alice/projects/devops_os with your actual path.

Restart Claude#

• Quit Claude completely (⌘Q on macOS, Alt+F4 on Windows)
• Reopen Claude Desktop
• You should see a wrench icon 🔧 at the bottom right of the chat

Step 3: Generate Your First Pipeline (2 minutes)#

Ask Claude#

In Claude Desktop, ask:

Generate a GitHub Actions CI/CD workflow for a Python Flask API with pytest tests and Docker build to Docker Hub.

Claude will:

1. Call the generate_github_actions_workflow tool
2. Return a ready-to-use YAML file
3. Explain each stage

Copy the Output#

Claude shows the generated workflow YAML. Copy it and save as .github/workflows/ci.yml in your project.

Step 4: Try More Prompts#

Jenkins Pipeline#

Generate a Jenkins Declarative Pipeline for a Java Spring Boot application with Maven build, SonarQube scan, and ArgoCD deployment to production.

Kubernetes#

Create Kubernetes manifests (Deployment + Service) for a Node.js API with 3 replicas, port 3000, and image myregistry/myapp:v1.0.

GitLab CI#

Generate a .gitlab-ci.yml for a Python project with lint, test, and Docker build stages.

SRE/Observability#

Generate SRE monitoring configs (Prometheus alert rules, Grafana dashboard, and SLO) for a microservice called “payment-api”.

Common Tasks#

Task: Multi-language Pipeline#

Generate GitHub Actions for a project with Python backend and React frontend. Include separate test stages and a combined Docker build.

Task: GitOps Deployment#

Create ArgoCD Application manifests for deploying a service from GitHub to a Kubernetes cluster, with auto-sync enabled.

Task: Hardening Baseline#

Generate Kyverno policies for a production Kubernetes cluster following CIS Kubernetes Benchmarks.

Task: Dev Container#

Scaffold a devcontainer for Python + Go + Node.js development with pre-installed tools.

Troubleshooting#

“Wrench icon shows error”#

1. Check the path in claude_desktop_config.json is correct:

   ls /path/you/configured/mcp_server/server.py

2. Verify Python can run the server:

   python -m mcp_server.server

   You should see output about MCP tools being registered. Press Ctrl+C to stop.

3. Restart Claude completely (not just refresh).

“Tool not found” error#

• Ensure you have the latest version: git pull
• Reinstall dependencies: pip install -r mcp_server/requirements.txt
• Restart Claude

Other Issues#

See MCP Setup & Configuration for detailed troubleshooting.

Next Steps#

What’s Next?#

Once you’re comfortable generating configs with Claude, check out:

• CLI Reference — Run the same generators from the command line
• GitHub Actions Deep Dive — Understand every stage
• Infrastructure Hardening — Generate security policies
• SRE Configuration — Set up observability

Questions?#

• 🐛 Report bugs: GitHub Issues
• 💬 Ask questions: GitHub Discussions