huawei-cloud-modelarts-notebook-management

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates the execution of cloud management tasks via the hcloud ModelArts CLI tool. It generates commands based on user intent for instance, lease, tag, image, and storage management.
  • [REMOTE_CODE_EXECUTION]: The installation guide instructs users to download and execute a shell script directly from support.huaweicloud.com using the curl | bash pattern. This is a standard installation procedure for the vendor's official CLI tool.
  • [EXTERNAL_DOWNLOADS]: The skill references and fetches installation scripts from Huawei Cloud's official documentation and support domains.
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it interpolates user-controlled strings—such as instance names, flavor IDs, and storage URIs—directly into shell commands and JSON input files.
  • Ingestion points: User-supplied values for parameters like {instance_name}, {flavor_id}, {uri}, and {mount_path} in references/cli-command-examples.md and references/known-issues.md.
  • Boundary markers: The skill explicitly requires the agent to prompt the user for confirmation before executing any write operations (Create, Update, Delete, Start, Stop, etc.), as stated in SKILL.md.
  • Capability inventory: The agent has the ability to perform full resource lifecycle management, including resource destruction and billing-related actions, via CLI and SDK execution.
  • Sanitization: There is no explicit instruction for shell escaping or sanitizing user-provided strings before they are used in CLI arguments or JSON payloads.
  • [DYNAMIC_EXECUTION]: Includes a Python SDK fallback code block using huaweicloudsdkmodelarts to perform operations if the CLI encounters bugs, involving the dynamic instantiation of service clients.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 07:05 AM