ClawHub
Back to skill

Security audit

智谱CogView模型图片生成

Security checks for vulnerabilities and agentic risk

Overview

This is a simple image-generation skill that uses Zhipu CogView as stated, with privacy caveats but no hidden or malicious behavior found.

Install only if you are comfortable using Zhipu AI for image generation. Do not include secrets, confidential business information, or sensitive personal data in prompts, because prompts and translated prompts may be sent to the external CogView API. Also verify that the referenced generate_image.ps1 script is available in the installed package or environment before relying on the skill.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

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 (3)

Vague Triggers

Medium
Confidence
93% confidence
Finding
The skill description says it should be used whenever the user wants AI-generated images, which is a broad activation condition that can cause the agent to invoke this skill in many contexts without sufficiently checking user intent, sensitivity of the prompt, or whether external transmission is appropriate. In this skill, that risk is amplified because prompts are sent to a third-party API service and may also be transformed automatically before submission.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The markdown explains that the tool will translate prompts and call the Zhipu CogView model, but it does not clearly warn users that their prompt content will be transmitted to an external AI provider. This creates a privacy and data-handling risk, especially if users include sensitive personal, confidential, or regulated information in image prompts under the assumption the request stays local.

Natural-Language Policy Violations

Low
Confidence
79% confidence
Finding
The description specifies support for automatically translating Chinese prompts into English, and the entire skill documentation is presented only in Chinese without indicating any user-selectable language or locale option. This may constitute a language/locale policy issue if the skill assumes a specific language environment rather than offering opt-in or alternatives.

Static analysis

No suspicious patterns detected.