azure-compute
Pass
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: SAFE
Full Analysis
- Administrative Command Templates: The skill provides structured workflows for managing Azure resources using the
azCLI. These tasks, such as resetting credentials, restarting VMs, and adjusting network security groups, are primary functions for cloud administration and are implemented using standard command-line patterns. - Official Resource Integration: The skill fetches data from
learn.microsoft.comandprices.azure.comto provide users with current pricing and technical guidance. These are official vendor resources, and their use is consistent with the skill's purpose. - Credential Safety Practices: Instructions for sensitive operations, such as password resets or SSH key updates, correctly use placeholders (e.g.,
<new-password>) rather than hardcoded secrets, following secure development practices. - Indirect Prompt Injection Surface: The process of fetching and summarizing external documentation presents a theoretical surface for indirect prompt injection. 1. Ingestion points: Remote content is fetched from
learn.microsoft.comin thevm-recommender.mdandvm-troubleshooter.mdfiles. 2. Boundary markers: Explicit markers are not defined in the markdown bodies. 3. Capability inventory: The agent is instructed to useazCLI tools for resource management. 4. Sanitization: External content is summarized for the user without explicit sanitization steps. This is considered a low-risk functional characteristic given the trusted source of the data.
Audit Metadata