ClawHub

Agent Self-Care

Security checks across malware telemetry and agentic risk

Overview

This skill is for agent maintenance, but it grants recurring authority to terminate agents/processes and write memory notes without enough scoping or approval controls.

Install only if you deliberately want autonomous agent maintenance. Prefer manual or dry-run operation first, require confirmation before terminating sub-agents or processes, restrict cleanup to resources clearly owned by this workflow, and review any cron schedule and memory files it creates.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Vague Triggers

Medium
Confidence
92% confidence
Finding
The skill is designed to trigger on very broad phrases such as "optimize", "cleanup", and "health check", which can cause it to run in contexts where the user did not intend autonomous maintenance actions. Because the skill includes destructive behaviors like killing sub-agents/processes and writing files, accidental invocation can lead to unintended service disruption or state changes.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The workflow instructs the agent to kill stale sub-agents and hanging processes using heuristic criteria, but it does not require explicit approval, dry-run output, or safety validation before terminating resources. In a self-maintenance skill scheduled to run automatically every 5 minutes, this raises the risk of disrupting legitimate long-running work or causing denial of service through over-aggressive cleanup.

Missing User Warnings

Low
Confidence
84% confidence
Finding
The skill directs the agent to document improvements in persistent files under memory/daily without warning that operational data, prompts, or sensitive task context may be retained across runs. Even if intended for benign retrospectives, silent persistence can create privacy, data minimization, and prompt-leakage risks over time.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal