The Agent Skills Directory

[Command Generation]: The skill dynamically constructs Azure CLI commands and REST API calls using user-provided values for parameters such as model names, SKUs, and capacity settings. While this enables full customization of the deployment, it represents a surface where user inputs are interpolated into executable commands.
[Metadata Access]: In the workflow for Anthropic models, the skill retrieves tenant metadata, including the country code and organization name, via the Azure Management API. This information is used to populate the modelProviderData field in the deployment payload, as required by the vendor's specific implementation.
[Dynamic Script Execution]: The workflow involves generating Bash and PowerShell script templates at runtime based on user selections. These scripts are then executed to perform resource management tasks, which is a common pattern for administrative tools but relies on the integrity of the input and the agent's execution environment.
[Input Validation]: The skill implements some sanitization, such as a regular expression check for deployment names (^[\w.-]{2,64}$), and guides users toward selecting values from system-queried lists for model versions and SKUs to mitigate potential configuration errors.

customize