The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill manages Azure Kubernetes Service clusters using user-provided parameters (such as resource group names, cluster names, and node pool configurations) and executes them via az, kubectl, and helm commands. This creates a surface where malicious input could potentially influence infrastructure operations if the agent does not properly validate parameters.
Ingestion points: User-supplied values for cluster configuration and management commands in SKILL.md.
Boundary markers: None identified.
Capability inventory: Subprocess calls for az, kubectl, and helm are described throughout the documentation.
Sanitization: No explicit sanitization or parameter validation logic is provided within the instructions.
[METADATA_POISONING]: The skill includes metadata and prose with dates and events set in the future (e.g., date_added: '2026-09-20' and claims of project retirement in 2026). Additionally, the YAML frontmatter includes a risk: critical label which contradicts the observed functional behavior and standard usage patterns of the documented tools.
[EXTERNAL_DOWNLOADS]: Fetches Helm charts and repository information from well-known community sources, including the Kubernetes project and Prometheus community GitHub pages. It also utilizes the az aks install-cli command to download management tooling.
[COMMAND_EXECUTION]: Employs Azure CLI (az), kubectl, and helm to perform cluster management and deployment tasks.

azure-aks