Happy Horse 视频工具箱｜AI-HIVE

• Outcome: Pass
• Latest audit: Aug 25, 2026 · 6:48 AM UTC
• Version: 1.0.0

Overview

This skill is a disclosed AI-HIVE video workflow helper that uses user-provided API credentials and media files for video generation and local edits, with no hidden destructive or exfiltration behavior found.

Before installing, understand that this skill can store an AI-HIVE API key, upload media files you explicitly provide, submit potentially billable video generation tasks, and write generated outputs locally. Use it only with authorized media, review parameters before generation, and avoid storing the API key in shared environments.

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (5)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 91% confidence
• Finding: The skill embeds executable workflows that use environment variables, local files, shell commands, and network access, yet it declares no permissions or capability boundaries. This is dangerous because users and hosting platforms cannot accurately assess what the skill may access or modify, increasing the risk of unintended file access, secret exposure, or outbound requests to external services.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 94% confidence
• Finding: The documented behavior for a narrowly scoped Happy Horse video toolkit appears to extend into broader AI-HIVE capabilities such as general chat, image generation, account or wallet queries, arbitrary model enumeration, and local credential setup. This mismatch is dangerous because it expands the attack surface beyond user expectations, enabling access to account-sensitive information and generalized remote operations that are not necessary for the stated purpose.

Vague Triggers

Medium

• Confidence: 83% confidence
• Finding: Triggering the skill on broad phrases like 'AI视频' and '视频编辑' can cause accidental invocation for many unrelated requests. In a skill that can lead to network calls, file handling, and potentially billable generation tasks, overly broad activation increases the chance of unintended actions, user confusion, and misuse outside the intended Happy Horse workflow.

Vague Triggers

Medium

• Confidence: 85% confidence
• Finding: The invocation logic mixes keyword matching with broad business scenarios, making the boundaries of when the skill should activate unclear. This is dangerous in context because the skill supports external API interactions and media-processing workflows, so ambiguous routing may cause the wrong tool to engage, expose users to unintended cost, or process unauthorized media.

Natural-Language Policy Violations

Medium

• Confidence: 84% confidence
• Finding: The default prompt hard-codes Chinese output ('production-ready Chinese workflow') regardless of the user's language preference. This is not a classic security bug, but it can override user intent and create unsafe ambiguity if users cannot clearly review prompts, commands, or compliance-sensitive output in their preferred language.

Static analysis

No suspicious patterns detected.