Security audit

AliExpress 速卖通电商视频生成与编辑

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent AI Hive video-generation helper for AliExpress listing media, with expected cloud upload, API key, and local output behavior.

Install only if you are comfortable sending selected product prompts and media to AI Hive. Use a dedicated API key where possible, review product and music rights before upload, and remember generated outputs are saved locally unless you use no-download mode.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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 (5)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 92% confidence
Finding: The skill invokes shell commands, reads/writes local files, accesses environment variables, and performs network operations, yet no permissions are declared to inform or constrain those capabilities. This creates a transparency and trust-boundary problem: users may invoke a seemingly simple ecommerce video skill without realizing it can install packages, contact external services, and handle local media or secrets.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 96% confidence
Finding: The documented purpose is narrowly scoped to AliExpress product-video generation, but the underlying behavior apparently includes broader AI Hive capabilities such as generic text chat, image generation, model enumeration, media upload, and account or wallet queries. That scope expansion increases the attack surface and can enable unintended data exposure, account reconnaissance, or off-purpose use that users did not meaningfully consent to.

Description-Behavior Mismatch

Medium

Confidence: 94% confidence
Finding: The file contains a full general-purpose AI Hive client for text chat, image generation, model listing, uploads, and account queries, while the declared skill is specifically for AliExpress video generation/editing. This scope expansion increases the attack surface and enables users or upstream agents to invoke unrelated capabilities that may bypass policy or product constraints intended for this skill.

Context-Inappropriate Capability

Low

Confidence: 88% confidence
Finding: Exposing a user-info endpoint gives this skill an account introspection capability unrelated to creating or editing AliExpress videos. Even if limited to the current account, it can reveal billing, balance, or identity metadata that need not be accessible from a narrowly scoped media skill.

Context-Inappropriate Capability

Low

Confidence: 84% confidence
Finding: Listing all available models exposes broader platform capabilities than necessary for a fixed-purpose ecommerce video skill. This can aid capability discovery and make it easier to repurpose the skill beyond its stated scope, especially when combined with the generic chat/image/video operations present elsewhere in the file.

Static analysis

No suspicious patterns detected.