ClawHub
Back to skill

Security audit

GPT Image 2.6 电商详情图生成 | 卖点图 | 详情页长图 | 参数信息图

Security checks for vulnerabilities and agentic risk

Overview

The skill provides a coherent image-generation workflow with disclosed uploads and spending confirmation, though its CLI setup changes persist.

Review the global CLI installation or upgrade because it can affect other workflows. Use reference images you are comfortable sending to Qinghu, verify the configured service address, protect your API key, and check the quote before confirming generation. The review covers the complete bundled skill instructions, not the external CLI's implementation; publisher identity was not supplied in workspace metadata.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Automatically installs or upgrades a global CLI without warning about system-wide changes.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

The skill directs the agent to run npm i -g @iqinghu/qhkit@latest automatically when the CLI is missing or outdated, without obtaining approval for a persistent global environment change. This can replace a CLI used by other workflows and execute package lifecycle scripts with the agent's permissions; an unpinned release also introduces supply-chain risk. Installing the CLI is relevant to the image-generation task, but the later confirmation for generation charges does not authorize software installation.

Content

No source excerpt is available for this finding.