ClawHub
Back to skill

Security audit

3C产品结构与功能演示｜AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI-HIVE video-production helper with expected local media, API key, network, and ffmpeg use, and I found no hidden or destructive behavior.

Install only if you intend to use AI-HIVE for product video work. Treat the AI-HIVE API key as sensitive, review prompts and routing before paid generation, and upload only media you are authorized to use.

SkillSpector

Vulnerability Patterns
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding
The skill exposes code-capable behavior including environment-variable access, file read/write, shell execution, and network access, but does not declare permissions or clearly constrain those capabilities. This creates a trust and review gap: a caller may invoke the skill expecting a documentation-only workflow while the included commands and scripts can access local files, secrets, and external services, increasing the chance of unintended data exposure or unauthorized actions.

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding
The documented purpose is narrowly framed as 3C product feature/structure video production, but the implementation reportedly also supports generic text chat, generic image generation, account/wallet retrieval, broad model catalog access, and standalone ffmpeg operations. That mismatch is dangerous because it conceals materially broader capabilities than users and reviewers would expect, enabling misuse of account data, off-scope content generation, or local media manipulation under the cover of a narrowly described skill.

Natural-Language Policy Violations

Medium
Confidence
86% confidence
Finding
The default prompt hard-codes Chinese output regardless of the user's language preference, which can cause unintended behavior and reduce user control over how the agent responds. While this is not a classic security bug, it is a real prompt-safety and policy issue because it silently overrides user intent and could cause misunderstandings in workflows or generated commands.

Static analysis

No suspicious patterns detected.