ClawHub
Back to skill

Security audit

Comfy Story Video

Security checks across malware telemetry and agentic risk

Overview

The skill’s media-generation purpose is clear, but its script can turn normal theme/story text into unsafe local shell commands.

Install only if you trust the themes and story text passed to it, and prefer a patched version that replaces os.system with subprocess.run argument lists and sanitizes filename components. Keep ComfyUI bound to localhost and run the skill in a limited local workspace.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • 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)

os.system() or os exec-family call

High
Category
Dangerous Code Execution
Content
# 使用 say 命令生成语音
        cmd = f'say -v "Ting-Ting" -o "{temp_aiff}" "{text}"'
        os.system(cmd)
        
        # 转换为 mp3
        cmd = f'ffmpeg -i "{temp_aiff}" -codec:a libmp3lame -qscale:a 2 "{output_path}" -y 2>/dev/null'
Confidence
97% confidence
Finding
The code builds a shell command with untrusted text content and executes it via os.system. Because the story text is inserted directly inside a quoted shell string for the macOS say command, crafted input containing shell metacharacters or quote-breaking content could trigger command injection and arbitrary local command execution.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.