ClawHub
Back to skill

Security audit

Video Analyzer

Security checks for vulnerabilities and agentic risk

Overview

This skill coherently extracts video frames and metadata for user-directed visual analysis, with only a manual cleanup command that users should run carefully.

Before installing, understand that this skill may read user-selected video files or video directories and write extracted JPEG frames plus metadata to an output folder. Be careful with the cleanup command: verify the output directory exactly before deleting it.

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)

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding
This is a mismatch because the description says the skill analyzes video content using vision, but the code only extracts frames and metadata and then instructs the user to use a separate image tool manually. The advertised content-understanding behavior is not implemented here. Batch support and frame extraction do match the description, but the primary claimed analysis capability is overstated, and the code also includes metadata extraction not described.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The markdown instructs users to run `rm -rf <output_dir>`, which is a destructive deletion operation, but it provides no warning about the risk of accidental data loss if the path is wrong or expanded unexpectedly. For markdown files, safety-affecting behaviors should include a clear warning when they can impact user data or system integrity.

Static analysis

No suspicious patterns detected.