The Agent Skills Directory

Command Execution via Local Scripts: The skill relies on several Bash and PowerShell scripts to automate complex validation sequences. These scripts invoke multiple command-line tools, including the Azure CLI (az), Azure Developer CLI (azd), Terraform, and Docker. This is a standard approach for infrastructure-as-code validation, but it involves substantial interaction with the local system's shell and installed utilities.
Indirect Prompt Injection Surface: The validation process involves scanning and analyzing user-provided project files, such as C# source code, Terraform configurations, and Bicep templates. Since the skill acts based on the patterns it finds in these files (e.g., detecting Aspire components or secret storage settings), malformed or intentionally crafted data in the project files could theoretically influence the agent's workflow or findings. The skill mitigates this by using a deterministic, step-based state machine defined in its workflow scripts.
Dynamic Execution Pattern: A Python one-liner is used within one of the shell scripts to perform JSON syntax validation on configuration files. While this is a common and efficient method for data validation, it involves the dynamic execution of a script snippet at runtime.
Authorization and Configuration Guidance: The skill provides instructions that guide the agent through a specific validation process, using authoritative language to ensure consistency. These instructions include safety-conscious checks, such as requiring human approval for destructive actions and ensuring that deployment plans are approved before proceeding.

azure-validate