视觉素材美工虾

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:41 PM UTC
• Version:
  1.0.0

Overview

This image-generation skill is coherent and purpose-aligned, with expected local script execution, model/API use, and local output storage, though users should be careful with cloud generation and optional uploads.

Install only if you are comfortable running local Python image-generation code that may use an OpenAI API key, download/load ML models, and save generated images locally. Use non-sensitive prompts for cloud generation, restrict API keys where possible, and approve Feishu uploads only when you intend to share the output.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 85% confidence
• Finding: The skill advertises and documents behavior that uses external model providers and local execution capabilities, but no permissions are declared. This creates a transparency and policy-enforcement gap: users and the platform may not realize that prompts or generated content can leave the local environment or that the skill can access environment-backed credentials.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 93% confidence
• Finding: The documented behavior and declared purpose do not match the effective behavior: the skill may call third-party image APIs, download remote assets, and omit claimed output paths such as Feishu upload or structured JSON handling. This misrepresentation is dangerous because users may disclose sensitive prompts or assets under the assumption processing is local or outputs behave as described, leading to unexpected data exposure and trust bypass.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The skill states that outputs are saved locally and may be uploaded to Feishu cloud storage, but it does not clearly warn users about retention, visibility, third-party access, or privacy consequences. In a visual-generation workflow, prompts and generated images may contain sensitive business, personal, or copyrighted material, so silent storage or upload can cause unintended disclosure.

VirusTotal

66/66 vendors flagged this skill as clean.