ClawHub
Back to skill

Security audit

OpenClaw Security Analysis

Security checks across malware telemetry and agentic risk

Overview

This appears to be a local security-scanning skill, but it overstates some security capabilities and may print detected secrets into logs.

Install only if you understand its limits: treat dependency results as informational rather than a vulnerability audit, do not rely on it for JavaScript code scanning, and avoid running the secret detector in shared terminals or CI logs unless output is captured securely. Review any adjacent c-support library before using C/C++ scanning.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
96% confidence
Finding
The tool advertises dependency vulnerability checking, but the implementation only parses requirement lines and prints advice to use external tools. This creates a false sense of security: users may believe dependency risk was assessed when no actual vulnerability analysis occurred, causing known vulnerable packages to be missed.

Intent-Code Divergence

Medium
Confidence
91% confidence
Finding
The module docstring claims JavaScript support, but the scanner logic only analyzes Python and C/C++ code paths. In a security tool, unsupported-language claims are dangerous because users may rely on scans that never happened, leaving JavaScript vulnerabilities undetected.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The secret scanning command prints matching lines back to stdout, and the masking only covers some long quoted values. Secrets embedded in unquoted formats, shorter tokens, URLs, or partially matched data may still be exposed in terminal logs, CI output, or shared artifacts, turning detection into disclosure.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.