Security audit

Temu 电商视频生成与编辑

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI Hive video-generation helper for Temu product media, with some extra dormant client code that is not exposed by the packaged entrypoint.

Install only if you are comfortable giving this skill an AI Hive API key and uploading selected product images, videos, or audio to AI Hive for generation. Avoid passing confidential media unless that third-party processing is acceptable, review live costs before batch runs, and use the documented generate/task/upload/init commands rather than treating the script as a general AI client.

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: 93% confidence
Finding: The skill advertises executable behavior with access to environment variables, local files, shell, and network operations, but does not declare permissions or constraints. This creates a transparency and least-privilege problem: users and the platform cannot accurately assess what the skill may access or exfiltrate when invoked.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 97% confidence
Finding: The documented purpose is narrowly scoped to Temu ecommerce video generation/editing, but the underlying capability set appears to include broad AI chat, image generation, account/balance retrieval, model discovery, and generic media upload. This mismatch is dangerous because it can mislead users and reviewers about the true attack surface, enabling unexpected data access, broader external communication, and abuse beyond the declared business purpose.

Description-Behavior Mismatch

Medium

Confidence: 97% confidence
Finding: The file embeds a full general-purpose AI Hive client with chat, model listing, upload, image generation, and video generation rather than a narrowly scoped Temu video skill. In an agent-skill context, this scope mismatch weakens least privilege and allows operators or downstream tooling to invoke unrelated capabilities, increasing the chance of policy bypass or unintended data transfer beyond the declared purpose.

Context-Inappropriate Capability

Medium

Confidence: 96% confidence
Finding: The presence of a text chat command is outside the declared Temu video-generation/editing purpose and gives the skill a general LLM interface. In agent environments, extra unconstrained text channels can be abused to send arbitrary prompts and sensitive context to a third-party API, expanding the attack surface beyond what users expect from a video tool.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The image-generation command exceeds the advertised video-generation/editing scope and introduces another unrelated remote-generation capability. This broadening matters because it permits arbitrary media creation and file handling workflows that may bypass governance, cost controls, or user expectations tied to a Temu-specific video skill.

Static analysis

No suspicious patterns detected.