ClawHub
Back to skill

Security audit

LLM数据自动化

Security checks across malware telemetry and agentic risk

Overview

This is a simple instruction-only skill for generating Python/Pandas code, with no executable files, install hooks, credentials, persistence, or local data access in the artifacts.

Review generated Python or Pandas code before running it on important or sensitive data. The skill itself appears safe to install based on the artifacts, but generated code can still make mistakes or perform unwanted changes if run blindly; VirusTotal was pending and was not used as a negative signal.

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
90% confidence
Finding
The trigger description says the skill activates when a message contains the skill name or related instructions, which is overly broad and can cause unintended invocation. In an automation skill that generates Python and Pandas code, accidental activation increases the chance the system produces code in the wrong context, potentially leading to unsafe downstream execution or confusing cross-task behavior.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.