ClawHub
Back to skill

Security audit

GEO 运营助手

Security checks across malware telemetry and agentic risk

Overview

This is a coherent GEO content-optimization skill with disclosed optional webpage and content-analysis tools, but users should avoid feeding it sensitive URLs or private content.

Install if you need a GEO operations workflow. Use public or approved webpages, avoid sensitive internal URLs or confidential pasted content, and redact personal or proprietary details from validation screenshots and reports.

SkillSpector

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

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger conditions include broad phrases like requests for optimization or product/keyword improvement, which can cause the skill to activate outside a narrowly intended GEO context. Over-broad activation can route unrelated user inputs into a workflow that encourages URL analysis, content ingestion, and optional web fetching, increasing the chance of unnecessary data exposure or unintended tool use.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.