ClawHub
Back to skill

Security audit

Auto-Watchdog

Security checks across malware telemetry and agentic risk

Overview

This watchdog skill fits its stated monitoring purpose, but it asks for persistent, privileged, quiet self-healing powers that can restart services and clean up logs/files without clear boundaries.

Install only if you intentionally want a persistent system watchdog with authority to restart services and manage logs/files. Before enabling it, restrict it to named services and directories, require backups or retention for rotated logs, avoid hidden execution unless necessary, and confirm there is a clear way to disable, audit, and roll back its changes.

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
Findings (2)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill explicitly encourages silent automatic restarts, log rotation, and cleanup of services and files without requiring confirmation, rollback guidance, or warnings about service disruption and data loss. In an agent context, this can cause unintended modification of running infrastructure, hide operational failures, and destroy forensic evidence such as logs needed for debugging or incident response.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The setup instructions tell users to rotate logs and clean up files automatically when thresholds are met, but provide no safeguards on retention, target scope, or data preservation. This is dangerous because an automated agent could delete or truncate important files, disrupt active processes, or remove evidence needed to investigate failures or compromise.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal