Azure Cloud Migrate

microsoft/github-copilot-for-azureOfficial

Handles the full assessment and code conversion workflow when you're moving workloads from AWS, GCP, or Heroku into Azure. You point it at Lambda functions, Beanstalk apps, or Kubernetes manifests, and it generates a migration report, rewrites your code for Functions or Container Apps, and outputs everything to a separate directory without touching your source. The scenario matrix is comprehensive (Lambda to Functions, Fargate to Container Apps, even Spring Boot migrations), and it enforces a phase-locked workflow so you can't skip the assessment step. One thing to like is the explicit rule about scanning for hardcoded service URLs in Kubernetes apps, since those DNS names won't resolve in Azure. Hands off to azure-prepare when it's time to deploy.

GitHub stars: 217
last updated: Jun 11, 2026

View on GitHub

Install

$ npx -y skills add microsoft/github-copilot-for-azure --skill azure-cloud-migrate --agent claude-code

Installs into .claude/skills of the current project.

Files

Azure Cloud Migrate

This skill handles assessment and code migration of existing cloud workloads to Azure.

Rules

Follow phases sequentially — do not skip
Generate assessment before any code migration
Load the scenario reference and follow its rules
Use mcp_azure_mcp_get_azure_bestpractices and mcp_azure_mcp_documentation MCP tools
Use the latest supported runtime for the target service
Destructive actions require ask_user — functions global-rules | app-service global-rules
Report progress to user — During long-running operations (deployments, image pushes), provide resource-level status updates so the user is never left waiting without feedback — see workflow-details.md
Audit service discovery in app code — Kubernetes DNS names (e.g., http://order-service:3001) do not resolve in Container Apps. During assessment, scan source code for hardcoded hostnames/ports in HTTP clients and flag them for env-var-driven URL injection