ClawHub
Back to skill

Security audit

OpenClaw LLM Tools

Security checks across malware telemetry and agentic risk

Overview

This is a coherent local utility for defining and converting LLM tool schemas, with no evidence of hidden network access, persistence, credential use, or destructive behavior.

Install this if you need a local LLM tool-schema registry and converter. Run it only on files you choose, avoid copying the eval example, pin dependencies if reproducible installs matter, and add stronger validation plus explicit confirmations before exposing file-changing, API-changing, or other side-effectful handlers to an LLM.

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)

Missing User Warnings

Medium
Confidence
85% confidence
Finding
The skill advertises `write` and `exec`-capable tooling but does not include any user-facing warning about filesystem changes, command execution, or the risk of destructive actions. In an agent setting, normalizing these capabilities without guardrails can lead users or downstream agents to invoke dangerous operations without understanding their impact.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.