Back to skill

Security audit

Ultimate Frontend

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed ClawHub maintainer/developer skill set with powerful workflows, but they are purpose-aligned, documented, and user-directed.

Install only if you are working on ClawHub maintenance or development and understand that some workflows can run local commands, contact GitHub/Convex/ClawHub services, and perform staff moderation actions when authenticated. Use the safer autoreview opt-out flags if you do not want nested review to bypass sandbox prompts.

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.