ClawHub
Back to skill

Security audit

AI大模型专家｜TVC 广告片

Security checks across malware telemetry and agentic risk

Overview

This TVC video skill mostly matches its stated purpose, but it can implicitly submit AI-HIVE jobs and upload media using stored credentials without a clear confirmation step.

Install only if you are comfortable with the agent using your AI-HIVE API key to upload selected media, submit video-generation tasks, poll for results, and download outputs. Prefer explicit invocation and confirm cost, prompt, assets, and output directory before running generation, especially for batch or commercial work.

SkillSpector

By NVIDIA
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 (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
86% confidence
Finding
The skill declares no permissions, yet its documented behavior includes network access, shell execution, environment/config handling, and file read/write operations such as storing API keys and downloading outputs. This creates a transparency and least-privilege problem: users and reviewers cannot accurately assess what the skill can access or modify before use.

Tp4

High
Category
MCP Tool Poisoning
Confidence
93% confidence
Finding
The skill is presented as a TVC/video-generation tool, but the analysis indicates broader capabilities including chat/image endpoints, model enumeration, account and wallet retrieval, browser opening, and credential file writes. This mismatch is dangerous because users may authorize or invoke the skill expecting a narrow workflow while it can access account data and perform unrelated actions beyond the advertised scope.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The default_prompt instructs the agent to directly execute a TVC generation workflow, save a taskId, and retrieve results without stating narrow activation conditions or requiring explicit user confirmation. In combination with allow_implicit_invocation: true, this broad invocation text can cause the skill to trigger on loosely related requests and initiate external actions such as uploads, job submission, polling, and downloads unexpectedly.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.