The Agent Skills Directory

Infrastructure and Command Execution: The skill performs extensive cloud management tasks using the Azure CLI (az) and PowerShell. This includes creating resource groups, provisioning services like App Service and Container Apps, and executing application deployments. These actions are protected by multiple approval gates where the user must explicitly confirm the plan and costs before execution.
Secret Management and Data Protection: Security is a core focus, with the skill automatically generating secure passwords using openssl and storing them in Azure Key Vault. It enforces the use of system-assigned managed identities for service-to-service authentication, minimizing the need for hardcoded secrets. A dedicated 'Blocked Patterns' reference prevents the execution of destructive or unsafe commands.
Automated Validation and Review: Before any deployment, the skill conducts multi-layer self-reviews and conformance checks. This includes syntax validation with bicep build, security baseline checks, and a comparison against established Well-Architected Framework (WAF) patterns. It also uses a dedicated sub-agent to ensure generated infrastructure code adheres to safety and best practice standards.
External Resource Interaction: The skill interacts with official Azure and Microsoft services to fetch pricing data, verify quotas, and download deployment tools like the Static Web Apps CLI. All network operations and downloads are directed toward trusted vendor infrastructure and well-known services.
Workspace Integration: To tailor deployments, the skill analyzes local workspace artifacts such as manifests (package.json), configuration files, and environment variable templates. While this involves processing untrusted repository data, the risks are mitigated by the fact that the resulting architecture and deployment steps are presented to the user for final approval.

azure-app-onboard