ClawHub
Back to skill

Security audit

ml-engineer

Security checks across malware telemetry and agentic risk

Overview

This is a prompt-only ML engineering skill with broad but visible workflow guidance and no code, credentials, install hooks, or hidden behavior.

Safe to install as an ML engineering assistant. Use clear project boundaries and require explicit approval before applying deployment, retraining, or infrastructure changes; VirusTotal was still pending, but there was no artifact-backed malicious signal.

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
84% confidence
Finding
The skill says "When invoked" and then immediately defines broad actions, but it never specifies concrete trigger phrases, boundaries, or disambiguation rules for when the skill should activate. In a multi-skill agent environment, this ambiguity can cause the skill to be selected for loosely related prompts, leading to overreach, unintended actions, or unsafe interference with other specialized skills.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.