ClawHub
Back to skill

Security audit

Batch

Security checks across malware telemetry and agentic risk

Overview

This is a simple batch-processing guidance skill with disclosed exec use and no hidden installer, scripts, or data exfiltration behavior.

Install this only if you want an agent to help run user-directed batch tasks. Review commands before execution, keep destructive operations behind explicit confirmation, and expect small checkpoint or failure-log files such as failed.json when the skill is used.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Confidence
90% confidence
Finding
The trigger keywords listed for this skill are very broad and map to common user language such as 'items', 'progress', 'multiple', and 'process'. In an agent ecosystem, this can cause the skill to activate in unrelated contexts, increasing the chance that exec-capable instructions are injected into workflows where the user did not intend to use this batch-processing skill.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.