ClawHub
Back to skill

Security audit

Weiliaozi Skill

Security checks across malware telemetry and agentic risk

Overview

The available evidence shows a prompt-only dual-use analysis concern, but no artifact-backed evidence of hidden execution, data theft, persistence, or destructive behavior.

Before installing, review the flagged SKILL.md section and avoid using it for sabotage, disruption, deception, or resource-denial planning. Based on the supplied clean VirusTotal result and lack of artifact-backed execution or persistence concerns, this does not need a Review hold.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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)

Intent-Code Divergence

Medium
Confidence
91% confidence
Finding
The skill explicitly frames itself as non-directive, but then includes concrete adversarial sequencing such as disrupting plans, weakening resources, and preparing countermeasures. Even with disclaimer language, these prompts can steer the agent into generating operationally useful hostile strategy, especially in business, policy, or competitive contexts where a user can map them onto real targets.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.