ClawHub
Back to skill

Security audit

Landing Page Builder

Security checks across malware telemetry and agentic risk

Overview

This is a marketing-planning skill made of markdown guidance, with no code execution, credential access, persistence, or hidden data handling found.

Reasonable to install for ecommerce marketing and landing-page planning. Before publishing outputs, verify product claims, review counts, urgency/scarcity statements, endorsements, customer photo permissions, privacy disclosures, tracking-pixel use, and refund/guarantee language for truthfulness and legal compliance.

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 (1)

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill explicitly recommends using customer reviews and before/after images, including "real customer photos (with permission)," but it does not provide any guidance on obtaining, documenting, and verifying consent, handling rights to user-generated content, or avoiding disclosure of sensitive personal information. In a marketing workflow, this omission can lead users to reuse testimonials or identifiable images without adequate authorization, creating privacy, publicity-rights, and compliance risk.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.