淘宝 电商图片生成与编辑

• Outcome: Pass
• Latest audit: Aug 17, 2026 · 3:58 AM UTC
• Version: 1.0.0

Overview

This skill is an AI Hive image-generation helper for ecommerce product images, with expected network upload, API-key storage, and local result downloads disclosed in its artifacts.

Install only if you are comfortable sending the product reference files you choose to AI Hive and storing an AI Hive API key locally. Use the documented generate/task/init commands, review any file path before upload, and be aware that API use may incur provider charges.

SkillSpector

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 (6)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 90% confidence
• Finding: The skill advertises operational commands that invoke shell execution, environment-variable use, filesystem access, and network activity, but it does not declare corresponding permissions. This creates a transparency and policy-enforcement gap: users and the platform may treat the skill as low-risk image prompting while it can actually access local files, read secrets from the environment, and make outbound requests.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 97% confidence
• Finding: The declared purpose is narrowly scoped to Taobao/Tmall product image generation, but the referenced tooling apparently supports much broader AI Hive operations including text chat, video generation, uploading arbitrary media, model enumeration, wallet/user queries, and API key management. This mismatch is dangerous because it hides a general-purpose networked automation surface behind a benign image-editing description, increasing the risk of unauthorized data upload, unexpected charges, or misuse beyond user intent.

Description-Behavior Mismatch

Medium

• Confidence: 92% confidence
• Finding: The file is presented as a Taobao/Tmall image-generation skill, but it actually bundles a broader general-purpose AI client with text chat, model listing, media upload, and video generation. This scope mismatch is dangerous because users and orchestrators may grant trust, permissions, or policy exceptions based on the narrow ecommerce-image description while the code exposes materially broader capabilities.

Context-Inappropriate Capability

Medium

• Confidence: 95% confidence
• Finding: Video generation is outside the declared purpose of a Taobao ecommerce image-editing skill, yet the code includes a full video workflow with media uploads and task polling. In a capability-gated agent environment, this creates unauthorized functional expansion and can bypass policy or review expectations tied to the skill's narrower description.

Context-Inappropriate Capability

Low

• Confidence: 88% confidence
• Finding: The text chat capability is unrelated to the advertised image-generation/editing role and gives the skill a general LLM interface beyond its stated use case. While less severe than video, it still expands the attack surface and may enable prompt-based misuse or policy bypass where only image tooling was expected.

Intent-Code Divergence

Medium

• Confidence: 91% confidence
• Finding: The module docstring describes a generic AI capability invocation tool rather than a specialized Taobao image skill. This discrepancy increases risk because reviewers, users, and automated systems may misunderstand the true functionality and approve broader code under narrower branding.

Static analysis

No suspicious patterns detected.