ClawHub
Back to skill

Security audit

无脸带货视频生成器｜AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI-HIVE video-generation helper that uses local scripts, an API key, uploads user-selected media, and may incur generation costs, with no evidence of hidden exfiltration or destructive behavior.

Install only if you intend to use AI-HIVE for commerce video generation. Review prompts and parameters before running generation because tasks may cost money, only upload media you have rights to use, and keep the persisted API key in ~/.ai-hive/config.json protected or use an environment variable instead.

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
95% confidence
Finding
The skill advertises executable workflows that use environment variables, local file access, network calls, and shell tooling, but it declares no permissions. That creates a capability transparency gap: a caller or hosting platform may invoke the skill without understanding that it can read/write files, access API keys, contact external services, and run ffmpeg or scripts, increasing the chance of unsafe execution or privilege overreach.

Tp4

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding
The skill is presented as a specialized faceless-commerce video generator, but the described/available commands extend to generic chat, model enumeration, media upload, task inspection, wallet/account retrieval, and broad ffmpeg operations. This mismatch is dangerous because users and policy layers may trust the narrower stated purpose while the actual capability set enables broader data access, external communication, and potentially billable or privacy-sensitive actions outside the advertised scope.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The default prompt auto-invokes the skill with a broad phrase ('Use $faceless-commerce-video-ai-hive to turn my request...') without embedding clear scope or trigger constraints. Because implicit invocation is also enabled, unrelated or weakly related user requests could be steered into this skill, causing overreach, unintended API usage, or generation of marketing/video workflows the user did not explicitly request.

Natural-Language Policy Violations

Medium
Confidence
94% confidence
Finding
The default prompt mandates a 'production-ready Chinese workflow' regardless of the user's language preference. This can cause instruction mismatch, reduced transparency, and user confusion, especially if safety-critical constraints, pricing, consent limits, or platform-policy warnings are presented in a language the user did not request or understand.

Static analysis

No suspicious patterns detected.