mlv-operations-cli
Pass
Audited by Gen Agent Trust Hub on Aug 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- Command Execution via Azure CLI: The skill utilizes
az restto interact with official Microsoft Fabric APIs. This is a standard and recommended method for cloud resource management, though it involves executing shell commands based on user input intended for administrative tasks.\n- Automated Safety Gate Option: An 'eval mode' is included that allows for the bypass of human-in-the-loop confirmations when specific phrases like 'I pre-approve' are provided in the prompt. This feature facilitates automated workflows and testing while maintaining a default manual-confirmation posture for general use.\n- Resource Identifier Processing: The skill handles user-supplied identifiers to perform its operations, which involves interpolation into command templates.\n - Ingestion points: User prompts for workspace names, lakehouse IDs, and table names (SKILL.md).\n
- Boundary markers: The skill does not define specific instruction-isolation delimiters for these identifiers.\n
- Capability inventory: The skill uses
az restto perform create, read, update, and delete operations on refresh schedules and job instances.\n - Sanitization: Input validation and sanitization are primarily managed by the underlying Azure CLI and the Microsoft Fabric API services.\n- Skill Update Procedures: The instructions include a session-based check to verify the skill's version. This is a constructive practice to ensure the agent uses the most current functional and security updates provided by the vendor.
Audit Metadata