电商批量内容工厂｜AI-HIVE

• Outcome: Pass
• Latest audit: Aug 25, 2026 · 2:18 AM UTC
• Version: 1.0.0

Overview

The skill is a coherent AI-HIVE e-commerce image/video generation helper, with expected network, upload, download, and API-key handling that users should treat as billable and sensitive.

Before installing, treat this as a third-party AI-HIVE integration: only upload media you are allowed to share, confirm parameters before billable generation, and avoid using init unless you are comfortable storing the API key locally in ~/.ai-hive/config.json.

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: 89% confidence
• Finding: The skill advertises runnable scripts that use environment variables, local files, network access, and shell-adjacent tooling like ffmpeg, yet it declares no permissions. This creates a transparency and governance gap: users or orchestrators may invoke a skill with capabilities they did not knowingly approve, increasing the chance of unintended data access, outbound requests, or local side effects.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 92% confidence
• Finding: The documented purpose is narrowly framed as an e-commerce batch content workflow, but the underlying behavior reportedly includes broader chat access, user-info and wallet queries, model enumeration, and local API-key setup/storage. That mismatch is dangerous because it obscures sensitive operations involving credentials, account metadata, and potentially billable actions, undermining informed consent and least-privilege use.

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: The file is packaged as a specialized ecommerce batch content/video skill, but its implementation exposes broad generic AI-Hive functionality including chat, model enumeration, user-info access, and arbitrary media upload. This increases the reachable capability surface beyond the declared purpose, which can enable misuse, policy bypass, or unintended data access in environments that rely on the manifest to constrain what the skill can do.

Context-Inappropriate Capability

Low

• Confidence: 83% confidence
• Finding: The skill provides a user/account information endpoint even though that capability is not necessary for ecommerce batch video generation. Unnecessary account introspection expands the data exposure surface and may reveal wallet/billing or account metadata to callers who should only be able to generate content.

Context-Inappropriate Capability

Medium

• Confidence: 91% confidence
• Finding: The generic text chat capability is unrelated to the declared specialized ecommerce batch video/content-factory behavior. In an agent setting, this creates an undeclared secondary channel that can be used for arbitrary prompt execution, broader model access, or handling sensitive uploaded content outside the intended workflow.

Intent-Code Divergence

Medium

• Confidence: 88% confidence
• Finding: The top-level documentation describes a generic AI capability tool, which conflicts with the claimed specialized ecommerce skill scope. This mismatch is dangerous because defenders, orchestrators, or users may trust the declared scope while the code actually supports broader actions, undermining transparency and policy enforcement.

Static analysis

No suspicious patterns detected.