ClawHub
Back to skill

Security audit

Skill Release Lifecycle

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only release checklist skill with disclosed verification commands and no hidden executable behavior.

This is reasonable to install if you want a reusable checklist for publishing or improving OpenClaw skills. Before using the generated agent prompt, review the target slug and command list because it can lead your agent to patch a local SKILL.md and run ClawHub/OpenClaw verification commands.

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.