Get started with Azure Skills

In this quickstart, you prepare, validate, and deploy an application to Azure by using Azure Skills.

Prerequisites

• Azure Skills installed (Install and configure Azure Skills)
• Azure CLI authenticated (az login completed successfully)
• GitHub Copilot CLI or Claude Code ready to use
• A sample application (or create a Node.js app)

Scenario

You have a Node.js application. You want to deploy it to Azure with:

• Web application hosting (Azure App Service)
• A storage account for application data
• Monitoring with Application Insights

Prepare your application

In your AI assistant, go to your project directory and ask:

Prepare my application for Azure deployment

The azure-prepare skill:

1. Analyzes your codebase
2. Identifies technology stack (Node.js, npm, and so on)
3. Creates .azure/plan.md with a deployment strategy
4. Generates infrastructure as code
5. Waits for your approval

Review the generated plan

Open .azure/plan.md and review:

• Project Information—Application name and deployment mode
• Requirements—Classification and scale (small, medium, large)
• Components—Technologies detected
• Deployment Strategy—Technology used to deploy your application (Azure Developer CLI, Bicep, Terraform, or Azure CLI).
• Architecture—Azure services selected
• Implementation Plan—Step-by-step tasks

Example plan content:

# Azure Deployment Plan

## Project Information
- Application: my-app
- Mode: NEW

## Requirements
- Classification: Web Application
- Scale: Small
- Environment: Production

## Components
- Runtime: Node.js 18+
- Package Manager: NPM
- Application Type: Express web server

## Recipe
- Type: AZD (Azure Developer CLI)

## Azure Services
- Azure App Service (web app hosting)
- Azure Storage Account (data)
- Application Insights (monitoring)

## Status: Awaiting Approval

Approve the plan

If the plan looks correct, tell your AI assistant:

Approve this plan and proceed to validation

The skill updates the plan status to Approved and moves to the next step.

Validate the deployment plan

Your AI assistant runs the azure-validate skill to check:

• Azure CLI access and permissions
• Bicep or Terraform template syntax, if applicable
• Azure subscription and region availability
• Service quota limits
• Required permissions for your account

Validation finishes and records proof of all checks in the plan. The plan status updates to Validated.

Review validation results

Check .azure/plan.md for the Validation Proof section, which shows:

• Commands executed
• Timestamp
• Results (passed or failed)

Example:

## Validation Proof
- Command: azd provision --preview
- Timestamp: 2026-03-16T14:22:00Z
- Result: ✓ All validation checks passed

If validation fails, review errors and ask your AI assistant to fix issues:

Fix the validation errors and try again

Deploy to Azure

When your plan is validated, tell your AI assistant:

Deploy my application to Azure

The azure-deploy skill:

1. Confirms plan status is Validated
2. Provisions Azure resources (storage, app service, monitoring)
3. Deploys your application code
4. Configures application settings
5. Provides your application endpoint

Deployment typically takes 3-5 minutes.

View your deployed application

After successful deployment, your AI assistant provides:

• Application URL (for App Service)
• Storage account name and access keys
• Application Insights instrumentation key

Example:

Deployment complete! 

Your app is live at: https://my-app-abcd1234.azurewebsites.net

Resources deployed:
- App Service: my-app-prod
- Storage Account: mystorageabcd1234
- Application Insights: my-app-insights

Monitor your app: https://portal.azure.com/...

To verify that your app is running, visit your application URL in a browser.

Verify your deployment

Test your application:

1. Visit your URL—Open the application URL in a browser.
2. Check monitoring—View logs in Application Insights.
3. Test functionality—Use key features of your app.

Ask your AI assistant for monitoring status:

Show me the application logs and performance metrics

Your AI assistant queries Application Insights and displays recent activity, errors, and performance data.

Update and redeploy

If you make code changes, redeploy easily:

1. Update your code in your editor.

2. Ask your AI assistant:

   Update the deployment with my latest changes
   

3. The skill runs azure-prepare to check for changes, then azure-deploy to update your resources.

Updated resources reuse existing infrastructure. Only changed components redeploy.

Clean up resources

When you no longer need your application, delete Azure resources to avoid charges:

Delete all Azure resources for this application

Your AI assistant:

1. Lists resources to be deleted (for your confirmation)
2. Deletes the resource group and all contents
3. Confirms cleanup complete

Example:

Resources to delete:
- Resource Group: my-app-rg
- All contained resources

Are you sure? (yes/no)

Type yes to confirm deletion.

Troubleshooting

Deployment fails with authentication error

Problem: Your AI assistant can't authenticate to Azure.

Solution: Reauthenticate by using az login and try again.

Plan validation fails

Problem: Azure Skills reports validation errors.

Solution: Ask your AI assistant to review and fix issues:

Why did validation fail? Fix the errors.

Application not accessible after deployment

Problem: The provided URL returns an error or times out.

Solution:

1. Verify the URL is correct.

2. Wait 1-2 minutes for DNS propagation.

3. Check Application Insights logs for errors.

   Show me recent errors in Application Insights
   

Next steps

• Azure Model Context Protocol (MCP) Server documentation—Deeper technical details