The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Azure CLI (az) to execute Resource Graph queries (KQL) and retrieve reservation details. These operations are essential for its functionality and are explicitly restricted to read-only actions to prevent accidental resource modification.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes metadata from Azure resources.
Ingestion points: Data is ingested through az graph query and MCP tools (compute_vm_list, storage_account_list) which retrieve resource names, tags, and properties.
Boundary markers: The instructions do not define explicit boundary markers or delimiters for the ingested Azure data.
Capability inventory: The skill can execute subprocesses via az CLI and uv run, and write analysis reports to the local Plans/ directory.
Sanitization: There is no evidence of data sanitization or validation of the retrieved resource properties before they are processed by the LLM.

AzureFinOps