ClawHub
Back to skill

Security audit

Code Runner Local

Security checks across malware telemetry and agentic risk

Overview

This skill openly functions as a local multi-language code runner, but any code it runs has the same access as the user running it.

Install this only if you want an agent to run code on your machine. Treat every snippet as trusted software: review it first, avoid running code from unknown sources, and use a disposable container or sandbox for anything that might read files, use the network, run shell commands, or change the system.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (2)

Vague Triggers

Medium
Confidence
88% confidence
Finding
The skill description and usage guidance are broad enough that an agent may invoke this skill for general coding help, not just explicit execution requests. In this context, over-broad routing is risky because the skill is specifically designed to execute arbitrary code across many languages, which can turn ordinary programming prompts into unintended code execution opportunities.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
This skill intentionally executes arbitrary user-supplied code across many languages by writing it to temporary files and invoking interpreters/compilers through subprocesses. In an agent skill context, that is dangerous unless it is strongly sandboxed, because untrusted code can read files, access secrets, make network calls, spawn processes, or attempt host escape; the current script shows no isolation beyond a timeout and basic temp-file cleanup.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.dangerous_exec

Shell command execution detected (child_process).

Critical
Code
suspicious.dangerous_exec
Location
scripts/run-code.cjs:140