ClawHub
Back to skill

Security audit

double6-workbench-builder

Security checks across malware telemetry and agentic risk

Overview

This skill is a local offline workbench builder with disclosed file creation and browser-check steps, and I did not find hidden exfiltration, persistence, or destructive behavior.

Install only if you are comfortable with a local Python skill creating files under a run directory and launching a local browser for evaluation. Set DOUBLE6_BROWSER_PYTHON and DOUBLE6_CHROMIUM_EXECUTABLE only to trusted local binaries, and avoid entering real child, student, customer, financial, or health data unless you accept the skill's local-only storage model and export/clear controls.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • 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
Findings (2)

subprocess module call

Medium
Category
Dangerous Code Execution
Content
with tempfile.TemporaryDirectory() as temp:
                output_dir = Path(temp)
                collector = PACKAGE_ROOT / "scripts" / "double6_browser_collector.py"
                completed = subprocess.run([
                    str(browser_python), "-B", str(collector), "--run", str(run_dir),
                    "--output-dir", str(output_dir), "--chromium", str(chromium),
                ], capture_output=True, text=True, check=False)
Confidence
82% confidence
Finding
completed = subprocess.run([ str(browser_python), "-B", str(collector), "--run", str(run_dir), "--output-dir", str(output_dir), "--chromium", st

Lp3

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding
The skill instructs the agent to execute local Python CLI commands, read multiple reference files, write run artifacts, and potentially perform environment repair, but the manifest does not declare corresponding permissions or capabilities. This creates a trust and review gap: a host may treat the skill as low-privilege while it actually requires shell, file, environment, and possibly network access through Python tooling or dependency resolution.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.