ClawHub
Back to skill

Security audit

Tech Debt Ledger

Security checks across malware telemetry and agentic risk

Overview

The skill bundle is mostly coherent, but its autoreview helper defaults to running nested review with sandbox and approval bypass and may send diffs to fallback AI reviewers.

Install only in a ClawHub/Convex maintainer environment where these powers are expected. For private or untrusted code, run the autoreview helper with --no-yolo or AUTOREVIEW_YOLO=0, and set --fallback-reviewer none if diffs must not be sent to external AI tools. Restrict moderation and PR workflows to trusted staff credentials.

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.