ClawHub
Back to skill

Security audit

Video Compressor

Security checks for vulnerabilities and agentic risk

Overview

This skill mainly prints guidance for using a third-party video compressor, with some confusing legacy script names but no hidden upload, credential access, persistence, or destructive behavior found.

Before installing, note that this is a guidance skill, not a local compressor. Use the primary guide_video_compressor.py script rather than the legacy image/SVG-named scripts, and avoid uploading confidential videos to VideoAny unless you trust that service's privacy and retention practices.

SkillSpector

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)

Intent-Code Divergence

Medium
Confidence
87% confidence
Finding
This script claims to be a backward-compatible legacy entrypoint for image-to-SVG conversion, but it actually invokes video-compressor guidance instead. That mismatch can mislead users, automation, or upstream tooling into running an unrelated capability under a trusted legacy name, which is a supply-chain and operator-trust risk even though it does not itself execute obviously dangerous shell commands here.

Description-Behavior Mismatch

Medium
Confidence
90% confidence
Finding
Redirecting a script named convert_image_to_svg.py to video-compressor guidance is deceptive behavior because the filename and compatibility message imply preserved semantics that are no longer true. In a skill with shell execution permission, this kind of misdirection is more dangerous because callers may trust the entrypoint name and invoke broader downstream behavior than intended.

Static analysis

No suspicious patterns detected.