ClawHub
Back to skill

Security audit

Vib14 Skill

Security checks across malware telemetry and agentic risk

Overview

The skill bundle is coherent and purpose-aligned, but it includes powerful repo, GitHub, Convex, and moderation workflows that should be used only in trusted projects.

Install only in a trusted ClawHub/Convex development environment. Before using moderation, PR publishing, deploy, or autoreview commands, confirm the target repo/account and understand that some helper paths can access local diffs or send review prompts through configured external reviewer CLIs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.