ClawHub
Back to skill

Security audit

Voiceover Teaching Video Factory

Security checks for vulnerabilities and agentic risk

Overview

This skill is a local video-production workflow that uses expected media tools and includes clear limits against unrelated file access, credential access, network use, and automatic publishing.

Install only if you are comfortable giving the agent local access to the media files you provide and allowing it to run ffmpeg, ffprobe, Python helpers, and possibly a local browser for rendering. Do not authorize publishing or URL fetching unless you explicitly want that action and have checked rights for the source media and voiceover.

SkillSpector

Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (1)

Lp3

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding
The skill explicitly instructs the agent to read files, write to job and delivery directories, and invoke shell commands such as ffmpeg, ffprobe, Python, and bundled scripts, but no declared permissions are provided to bound those capabilities. Although the skill text includes safety-oriented scope restrictions, those are only prompt-level constraints and do not replace an enforceable permission model, so a host may grant broader access than intended.

Static analysis

No suspicious patterns detected.