on-demand-cluster
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
zillizCLI to interact with cloud resources. It defines several commands for creating, listing, describing, and deleting on-demand clusters. As a vendor-provided tool (zilliztech), this is expected behavior for cluster management. - [PROMPT_INJECTION]: The skill describes an interface where user-provided inputs (like project IDs and cluster names) are used as arguments for CLI commands. This presents a potential surface for indirect prompt injection or argument injection if the agent fails to sanitize user input. However, the documentation includes specific constraints for inputs (e.g., character limits and allowed character sets for cluster names), which provides a basis for validation.
- Ingestion points: Placeholders for command arguments in
SKILL.mdsuch as<project-id>and<cluster-name>. - Boundary markers: None explicitly defined.
- Capability inventory: Shell execution of the
zillizCLI. - Sanitization: The skill provides guidance on valid input formats (regex-like constraints for names and minimum values for integers), which helps mitigate accidental misuse.
Audit Metadata