eBay 电商视频生成与编辑

• Outcome: Pass
• Latest audit: Aug 17, 2026 · 5:38 AM UTC
• Version: 1.0.0

Overview

This skill is a disclosed AI Hive video-generation helper for eBay listing media, with some unused broader client code but no hidden or automatic unsafe behavior found.

Before installing, understand that selected listing photos or videos and prompts are sent to AI Hive for remote processing, and an AI Hive API key may be stored locally. Avoid passing confidential seller data unless you are comfortable with that service handling it.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• 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

Findings (7)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 83% confidence
• Finding: The skill advertises a narrowly scoped video-editing function, but its documented runtime invokes a Python script with shell execution, network access, and file read/write while declaring no permissions. That mismatch undermines least-privilege controls and can expose local files, environment secrets, or enable unintended outbound requests if the script or its dependencies are abused.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 93% confidence
• Finding: The stated purpose is conservative eBay listing video generation/editing, but the underlying behavior reportedly includes general AI chat, arbitrary model discovery, account and wallet retrieval, generic uploads, and broad AI Hive initialization. This scope expansion is dangerous because users may trust the skill with listing media while it can interact with unrelated account data and broader remote services not implied by the description.

Description-Behavior Mismatch

Medium

• Confidence: 92% confidence
• Finding: The file implements a generic AI Hive client with chat, image, video, upload, model listing, account-info, and task polling capabilities, which is materially broader than the declared eBay video-generation/editing purpose. This over-scoping increases attack surface and allows users or downstream agents to invoke unrelated capabilities under the trust boundary of a narrowly described skill.

Context-Inappropriate Capability

Low

• Confidence: 86% confidence
• Finding: Exposing user-info retrieval is outside the stated eBay listing video workflow and grants access to account metadata and potentially wallet/billing details. While not an exploit by itself, it creates unnecessary sensitive-data exposure within a skill that users would not expect to query account information.

Context-Inappropriate Capability

Medium

• Confidence: 90% confidence
• Finding: The skill exposes a general text-chat interface unrelated to eBay video creation/editing, enabling arbitrary prompt forwarding to external models. In the context of a narrowly scoped skill, this broad capability can be abused for data exfiltration, policy bypass, or unexpected charges, because users may trust it as a specialized media tool rather than a generic LLM proxy.

Context-Inappropriate Capability

Medium

• Confidence: 89% confidence
• Finding: Standalone image generation extends the skill beyond its declared video-editing/listing-video purpose and introduces additional external-generation capability and billing surface. This mismatch makes the skill harder to reason about and can enable misuse under a misleadingly narrow manifest.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The changelog states that the skill automatically uploads media, polls tasks, and downloads results, which implies transfer of user content to external systems. In a skill handling e-commerce listing videos that may contain proprietary product media or sensitive seller data, failing to clearly warn users about external data transfer creates a real transparency and privacy risk.

VirusTotal

64/64 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.