ClawHub

Ai Security Audit

Security checks across malware telemetry and agentic risk

Overview

This instruction-only security audit skill is not malware, but it asks for sensitive endpoint details and makes strong threat-intelligence claims without a clear lookup mechanism or privacy guardrails.

Review before installing. Only provide endpoint details you are comfortable sharing in an agent conversation, and treat breach, leaked credential, APT, or CVE results as advisory unless the agent shows exact source and query evidence. Run the suggested privileged server commands manually on the intended host only after checking placeholders, backups, and access impact.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger list is broad enough to activate on generic security-help requests, which can cause the skill to collect sensitive infrastructure details or steer users into this workflow when they may have intended a different, less data-collecting task. In a security-audit skill, overbroad activation increases the chance of unnecessary exposure of IPs, hostnames, provider, and region information.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill explicitly asks for public IP, hostname, cloud provider, region, and related network metadata without warning that these are sensitive operational details. Collecting and centralizing this information in a conversational workflow can create privacy and targeting risk, especially because the skill focuses on externally exposed AI services.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal