Amazon 亚马逊 电商图片生成与编辑

• Outcome: Pass
• Latest audit: Aug 17, 2026 · 6:03 AM UTC
• Version: 1.0.0

Overview

This is a disclosed AI Hive image-generation skill for Amazon seller assets, with user-selected uploads and API-key setup; unused generic helper code is present but not exposed by the active skill CLI.

Install only if you are comfortable sending selected product, packaging, and brand assets to AI Hive and storing an AI Hive API key locally. Review uploaded files before invoking the skill, especially because the upload helper does not enforce image-only extensions.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• 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
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (7)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 92% confidence
• Finding: The skill advertises no declared permissions, yet its documented execution path invokes Python scripts that can access the environment, filesystem, shell, and network. This creates a trust boundary problem: users and reviewers may assume a narrow image-editing capability while the skill can perform broader actions, increasing the chance of unexpected data access, exfiltration, or command execution if the backing script is unsafe or compromised.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 97% confidence
• Finding: The documented behavior suggests a focused Amazon image-generation skill, but the underlying capability set appears to include generic multimodal chat, video generation, account and wallet access, model/routing selection, arbitrary media upload, and browser-based API-key setup. This mismatch is dangerous because it can conceal materially broader privileges and data flows than users expect, enabling misuse of uploaded content, exposure of credentials, or unauthorized actions against external services.

Description-Behavior Mismatch

High

• Confidence: 96% confidence
• Finding: The file is a generic AI Hive client that exposes capabilities far beyond the declared Amazon ecommerce image-generation/editing skill, including chat, model enumeration, user-info access, video generation, and arbitrary media upload. In a skill context, this capability overreach violates least privilege and creates an opportunity for misuse, prompt-surface expansion, and access to account-scoped functions unrelated to the stated purpose.

Context-Inappropriate Capability

Medium

• Confidence: 91% confidence
• Finding: The skill can call a user-info endpoint and disclose account information and wallet balance even though that data is unnecessary for Amazon image generation. That expands the data-access surface and could expose sensitive billing or identity metadata to prompts, logs, or downstream consumers of the skill.

Context-Inappropriate Capability

Medium

• Confidence: 90% confidence
• Finding: The code supports arbitrary text chat with multimodal image input, which is unrelated to the advertised Amazon image-editing role. This broadens the skill into a general-purpose conversational interface, increasing risk of unauthorized data handling, policy bypass through alternate code paths, and misuse of the configured API key.

Context-Inappropriate Capability

High

• Confidence: 95% confidence
• Finding: The skill implements full video-generation workflows despite being presented as an Amazon image-generation/editing tool. That mismatch materially increases compute, billing, content-generation, and abuse surface, and allows consumers to invoke a more powerful modality than expected under the same trust boundary.

Intent-Code Divergence

Medium

• Confidence: 87% confidence
• Finding: The module docstring explicitly describes a generic AI capability invocation tool, which contradicts the specialized Amazon image-generation identity presented in metadata. This kind of capability/identity mismatch is dangerous because reviewers and users may grant trust based on a narrow purpose while the code actually supports broader operations.

VirusTotal

65/65 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.