The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents the use of the azd (Azure Developer CLI) and az (Azure CLI) tools to manage infrastructure and deployments. It specifically details the use of lifecycle hooks (preprovision, postprovision, postdeploy) within azure.yaml that execute shell commands for tasks like role assignments and post-deployment checks.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where it processes project-specific configuration files such as azure.yaml, Bicep files, and Dockerfiles. If these files were sourced from an untrusted third party, they could contain malicious configurations or shell commands intended to influence the agent's actions during the deployment pipeline.
Ingestion points: Project files (azure.yaml, Bicep templates, main.parameters.json).
Boundary markers: The skill instructs the agent to "Prepare the exact diff, target, estimated impact and rollback plan; execute only when the user explicitly authorizes that deployment."
Capability inventory: The skill uses shell execution via azd hooks and cloud resource management via az commands.
Sanitization: The skill recommends inspecting "redacted service logs" and warns against printing credentials into shared logs.
[EXTERNAL_DOWNLOADS]: The skill references official Microsoft documentation for the Azure Developer CLI and Container Apps. These references target trusted domains (learn.microsoft.com) and are used for informational purposes to guide the user on API and command verification.

azd-deployment