ClawHub
Back to skill

Security audit

huawei-cloud-modelarts-notebook-management

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent Huawei Cloud notebook management guide, but it asks for high-impact cloud authority and includes under-scoped safety guidance around installation, credentials, and mutating operations.

Review before installing if you want an agent to manage Huawei Cloud resources. Use least-privilege IAM, prefer read-only permissions unless you need lifecycle management, confirm the exact region and resource ID before any write action, avoid the pipe-to-bash install command, and keep AK/SK credentials in protected configuration or secret storage rather than prompts, scripts, or logs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger list includes broad English phrases such as "manage notebook", "notebook management", and "ModelArts notebook management", which can match ordinary user requests that are not explicit requests to invoke this skill. Because this skill enables lifecycle operations on cloud notebook resources, over-broad activation increases the chance of unintended tool selection and potentially destructive follow-on actions if a user confirms or the agent misinterprets intent.

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The acceptance criteria require confirmation for write operations in general, but they do not explicitly require user-facing warnings for non-delete destructive or service-impacting actions such as stopping instances, updating properties, detaching storage, or deleting images/image groups with downstream effects. In an infrastructure-management skill for notebooks, these omissions can lead to accidental disruption, data loss, or breakage of dependent workflows because operators may not be adequately informed of side effects before approving an action.

Missing User Warnings

Medium
Confidence
98% confidence
Finding
The guide instructs users to execute code fetched directly from a remote URL via `curl ... | bash`, which bypasses any opportunity to inspect the script before execution and creates a supply-chain execution path. If the remote content, transport, hosting account, or upstream page is compromised, users could run arbitrary commands on their local machine with their own privileges, potentially exposing Huawei Cloud credentials configured afterward or already present in the environment.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The SDK fallback example shows direct use of AK/SK and project_id placeholders without any warning about secure credential handling, storage, or leakage risks. In an agent skill context, users may copy this pattern into logs, scripts, shell history, or prompts, increasing the chance of credential exposure and subsequent unauthorized access to Huawei Cloud resources.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.