OpenClaw Security Suite

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:18 PM UTC
• Version:
  0.2.3

Overview

This is a coherent security-scanning skill, with user-directed file reading and LLM review that are expected for its purpose but should be used carefully on sensitive files.

Install if you want a local OpenClaw skill scanner, but treat results as advisory. Only run the AI review action on files you are comfortable sending to the configured LLM, and be aware the static directory scan appears to cover only top-level .ts and .js files rather than recursively scanning every nested file.

SkillSpector

By NVIDIA

Vulnerability Patterns

• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 88% confidence
• Finding: The skill metadata describes a scanner/reviewer, but the observed behavior set includes code execution via node:vm, signature verification, and runtime blocking logic that are not disclosed in SKILL.md. For a security-focused skill, undocumented execution and enforcement capabilities materially change the trust model and can expose users to unexpected code-processing or denial behaviors.

VirusTotal

63/63 vendors flagged this skill as clean.