ClawHub
Back to skill

Security audit

Agent Media: Generate AI-powered videos and images from the terminal using the `agent-media` CLI.

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent guide for using an AI video CLI, but users should be careful because provided media files and product URLs may be sent to external services.

Before installing, verify the npm package and GitHub project for agent-media-cli. Use only the account you intend to charge or associate with generated videos, and provide only product screenshots, local files, face photos, voice samples, URLs, and videos that you are comfortable uploading to the service.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill explicitly instructs the agent to visit a user-provided product site and extract image URLs, which causes external network access and may disclose browsing metadata such as IP address, user agent, timing, and the fact that the agent is interacting with that site. Because this happens as part of the prescribed workflow and without a clear consent or warning step, users may unintentionally trigger outbound requests to third-party domains.

Missing User Warnings

High
Confidence
97% confidence
Finding
The skill states that local file paths passed to `--broll-images` are auto-uploaded, but it does not prominently warn users that local screenshots will leave the machine and be transmitted to an external service. This creates a meaningful data exfiltration risk because screenshots can contain sensitive product data, credentials, internal dashboards, or customer information, and the workflow encourages uploading them as normal usage.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.