ClawHub

glm-v-model

Security checks across malware telemetry and agentic risk

Overview

This skill appears to be a straightforward Zhipu GLM vision-model helper, but it sends provided images and prompts to Zhipu's API.

Install only if you intend to use Zhipu's GLM vision API. Provide ZHIPU_API_KEY deliberately, install zai-sdk from a trusted source, expect API quota or billing use, and avoid sending confidential images, documents, screenshots, or prompts unless Zhipu's data handling terms are acceptable.

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 (2)

Vague Triggers

Medium
Confidence
87% confidence
Finding
The trigger description contains broad natural-language phrases such as 图片理解、图像识别、多模态分析、看图说话、图表分析、视频理解, which could cause the skill to be invoked in situations beyond the author's intent. Over-broad routing can expose user-provided images, videos, or prompts to an external model service unexpectedly, increasing privacy and data-handling risk.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The function sends user-provided images and prompt text to a third-party Zhipu API, which is a real data-exposure risk when inputs may contain sensitive content. In a vision-analysis skill, this behavior is expected, but the lack of visible disclosure, consent flow, or data-classification checks means users or integrators may unknowingly transmit private images, documents, or screenshots off-platform.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal