ClawHub
Back to skill

Security audit

Translate Book

Security checks across malware telemetry and agentic risk

Overview

This skill performs a disclosed book-translation pipeline and its command use, file writes, sub-agents, and cleanup behavior fit that purpose.

Install only if you are comfortable giving the skill access to the book file you name and allowing it to create a temp workspace, run local Calibre/Pandoc tools, and send chunks through sub-agents for translation. Keep intermediates if you need audit/recovery files, and use trusted Calibre/Pandoc binaries in your PATH.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

subprocess module call

Medium
Category
Dangerous Code Execution
Content
for path in possible_paths:
        try:
            result = subprocess.run([path, "--version"], capture_output=True, text=True, timeout=10)
            if result.returncode == 0:
                print(f"✓ Found Calibre ebook-convert: {path}")
                return path
Confidence
84% confidence
Finding
result = subprocess.run([path, "--version"], capture_output=True, text=True, timeout=10)

Missing User Warnings

Medium
Confidence
77% confidence
Finding
When --cleanup is supplied, the script deletes chunk and intermediate markdown files without a strong user-facing warning or safety checks beyond trusting temp_dir. In a translation workflow, those intermediates may be valuable audit/recovery artifacts, so accidental invocation can cause irreversible data loss and make troubleshooting or provenance review impossible.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal