What problem does it solve? Building and deploying a GitHub Copilot SDK application on Azure involves many moving parts: scaffolding the project, wiring authentication, configuring models, and provisioning Bicep infrastructure. This Skill routes you through the correct path for your scenario so you avoid misconfigurations and rework. ## Core Features & Use Cases - Scenario Routing: Directs you to the right workflow whether you are scaffolding a greenfield project, adding an SDK service to an existing repo, deploying an existing SDK app, or integrating the SDK into existing code. - Model Configuration: Supports three model paths including GitHub default models, specific GitHub models via listModels(), and Azure BYOM (Bring Your Own Model) with DefaultAzureCredential or ManagedIdentityCredential. - Deployment Pipeline: Orchestrates azure-prepare, azure-validate, and azure-deploy steps using the azure-samples/copilot-sdk-service azd template with Bicep infra and Key Vault token storage. - Use Case: You have an existing Express app and want to add a Copilot-powered chat endpoint hosted on Azure Container Apps with your own Azure OpenAI o4-mini deployment. The Skill guides SDK integration, BYOM provider config, and infra provisioning. ## Quick Start Ask the agent to scaffold a new Copilot SDK app with azd init using the azure-samples/copilot-sdk-service template and deploy it to Azure.