agent-bom discover azure

This Skill addresses the challenge of uniformly inventorying AI Agent and MCP-related assets in the Azure cloud environment. In practice, operations engineers and AI agent developers often need to quickly understand resources such as Azure OpenAI, Container Apps, AKS, Functions, and ML. However, manual queries are inefficient and prone to omissions, while granting long-term credentials to third-party tools poses security risks.

Usage is straightforward. You only need to ensure Python 3.11+ and the agent-bom tool are installed locally, and you possess Azure read-only credentials (via Azure CLI, workload identity, managed identity, or a service principal). You can then trigger the discovery process using natural language instructions, such as requesting to "scan all AI assets under the current Azure subscription and generate an inventory." The Skill will invoke the Azure SDK for read-only discovery, output a JSON file compliant with the agent-bom specification, and by default, it will remain in the discovery phase without automatically uploading or sharing the results.

It is particularly suitable for operations engineers and AI agent development teams that need to regularly inventory their Azure AI infrastructure. Organizations already using the agent-bom ecosystem for asset management can seamlessly integrate it into their existing workflows. For teams requiring audit or compliance evidence, this Skill provides a standardized inventory record, eliminating errors from manual documentation.

It is recommended to log in via Azure CLI and confirm you have at least Reader permissions before first use. Note that it is primarily focused on the discovery phase; scanning and subsequent analysis require the operator to explicitly run the agent-bom scan command. Additionally, it does not modify any Azure resources, nor does it persist credentials; all credentials exist solely within the operator's environment.