ClawHub
Back to skill

Security audit

AI大模型专家｜淘宝 电商图片生成与编辑

Security checks across malware telemetry and agentic risk

Overview

The skill mostly matches an AI-HIVE ecommerce image workflow, but it can be invoked implicitly to upload media and start paid external generation jobs without a strong trigger or confirmation boundary.

Install only if you are comfortable with AI-HIVE receiving prompts and any reference images you provide, with the skill storing an API key in ~/.ai-hive/config.json, and with the possibility that implicit invocation may start an external generation task. Review the API base URL and costs before use, and prefer explicit confirmation before uploading assets or submitting jobs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill documentation describes capabilities that require environment access, file read/write, network access, and shell execution, but it declares no permissions. This creates a transparency and trust problem: a user or hosting platform may invoke the skill without understanding that it can read local files, persist secrets, call remote APIs, and launch commands, increasing the risk of unintended data exposure or unsafe execution paths.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding
The skill is presented as a Taobao e-commerce image generation/editing tool, but the analysis indicates it also supports broader chat, video, audio, account, wallet, and model-discovery functions. This hidden scope expansion is dangerous because users may provide credentials, media, or account access under a narrower trust assumption, while the skill can access additional data types and operations not clearly bounded by its stated purpose.

Vague Triggers

Medium
Confidence
93% confidence
Finding
The skill enables allow_implicit_invocation without any visible trigger constraints, allowing the agent to invoke a real external image-generation workflow automatically. Because this skill uploads reference images, submits jobs, polls status, and downloads outputs through AI-HIVE, unintended invocation could cause unreviewed data transfer, unexpected actions, and unnecessary spend.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.