ClawHub
Back to skill

Security audit

Lazada 电商图片生成与编辑

Security checks for vulnerabilities and agentic risk

Overview

The skill is an API-backed Lazada image workflow that uploads user-selected assets and stores an AI Hive API key, with normal commands aligned to image generation.

Install only if you are comfortable sending selected product/reference images and prompts to AI Hive and storing an AI Hive API key locally. Use the provided generate/task/init commands, avoid uploading unrelated private files, and check AI Hive pricing before batch runs.

SkillSpector

Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding
The skill exposes effective capabilities to read/write files, access environment variables, use the network, and invoke shell commands, but does not declare those permissions. This creates a transparency and trust gap: users may authorize an image-editing skill without understanding it can access broader local and remote resources, increasing the chance of unintended data exposure or command execution through the supporting scripts.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding
The documented purpose is narrowly framed as Lazada ecommerce image generation/editing, but the underlying behavior reportedly supports broad multimodal chat, video generation, account info retrieval, generic uploads, model enumeration, and interactive API-key/browser configuration. This mismatch is dangerous because it can mislead users into granting access or providing sensitive assets under a narrow trust assumption while the skill can perform materially broader actions, including external data transfer and credential handling.

Static analysis

No suspicious patterns detected.