CAIRA

Use this skill when a user wants to build an AI, agent-based solution on the Azure AI platform (for example with Azure AI Foundry) using CAIRA and needs current context plus architecture decisions.

Best-fit scenarios

• Projects that want to use Azure AI platform capabilities, including Azure AI Foundry.
• Teams building agent-based applications that need CAIRA-guided decisions across infrastructure, platform, and application layers.

Core rules

• Treat the CAIRA repository as the single source of truth.
• Do not hardcode architecture names, module names, or fixed mappings.
• Discover what exists at runtime from repository/API listings, then decide.
• Support multiple layers (not only IaC): infrastructure, platform capabilities, application integration, and agent workflow composition.
• Default to passwordless/tokenless authentication for Azure access.
• Prefer Managed Identity or Azure Default Credential for Azure AI Foundry access.
• Avoid API keys unless the user explicitly requests key-based authentication.
• Before deployment/scaffolding actions, ask whether the user wants to customize provided deployment variables or use defaults when available.