ClawHub
Back to skill

Security audit

数字人持物展示视频｜AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI-HIVE video-production helper that uses user-provided media, an AI-HIVE API key, network calls, and local ffmpeg processing for its stated avatar product-video workflow.

Install only if you intend to use AI-HIVE for avatar/product video generation. Expect selected media files to be uploaded to AI-HIVE when you run upload or generation commands, and confirm prompts, routing, budget, and asset rights before any billable generation. Protect the AI_HIVE_API_KEY or ~/.ai-hive/config.json credential file.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill advertises runnable scripts that use environment variables, local file access, shell tools, and outbound network calls, but it does not declare corresponding permissions. This creates a trust and review gap: operators may invoke the skill without understanding that it can read/write files, access API keys from the environment, call external services, and execute local tooling such as ffmpeg.

Tp4

High
Category
MCP Tool Poisoning
Confidence
89% confidence
Finding
The documented purpose is a narrowly scoped workflow for product-holding avatar videos, but the described behavior appears to expose broader capabilities such as generic chat, image generation, model/account queries, media upload utilities, and local ffmpeg editing. That mismatch can let a user or orchestrator grant trust and invoke the skill under a narrow use assumption while actually enabling wider data access, content generation, external transfer, and billable actions than expected.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The default prompt automatically steers requests to this skill using a broad phrase ('turn my request into...') without constraining the types of user input that should activate it. In combination with a generative workflow that can produce runnable commands, this raises the chance of unintended invocation for adjacent requests, causing the agent to generate external API actions or sensitive media-production steps the user did not explicitly request.

Vague Triggers

Medium
Confidence
94% confidence
Finding
Enabling implicit invocation without contextual narrowing allows the platform to auto-select this skill even when user intent is ambiguous. Because the skill can upload reference assets, query pricing, route jobs, and download generated outputs through AI-HIVE, unintended activation could trigger external-service operations, mishandle copyrighted or biometric-like inputs, or facilitate deceptive marketing content if guardrails are bypassed.

Static analysis

No suspicious patterns detected.