Back to skill

Security audit

toolbelt

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed Toolbelt setup and data-query integration that asks for consent before network calls, config writes, uploads, sharing, or persistence-sensitive actions.

Install only if you want agents to use Toolbelt as a shared, persistent workspace. Review consent prompts carefully before allowing account provisioning, MCP config writes, document uploads, recorded findings, or share URL creation, because uploaded or recorded data may persist and share URLs can grant namespace access.

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.