ClawHub
Back to skill

Security audit

Video Director

Security checks across malware telemetry and agentic risk

Overview

This skill is a local storyboard generator that appears safe, with the main caveat that it adds a ClawHub-branded ending by default.

Reasonable to install for storyboard generation. Review the final scene before publishing outputs, because the default ending promotes clawhub.ai; use a custom JSON ending if you want neutral or client-specific text. Run the Node script only against intended input and output files.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The script appends a default branded call-to-action and domain reference ('clawhub.ai') to every storyboard unless callers explicitly override it. That creates hidden output manipulation beyond the declared role of neutral scene planning, which can mislead downstream agents or users and turn the skill into an undisclosed promotional channel.

Intent-Code Divergence

Medium
Confidence
92% confidence
Finding
The file header states the output is storyboard data generated from the topic and points, but the implementation also injects unrelated branded ending content. This mismatch is dangerous because other agents or users may trust the script as a pure transformer while it silently adds marketing material not derived from the provided input.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.