ClawHub
Back to skill

Security audit

Product Page Audit

Security checks across malware telemetry and agentic risk

Overview

This appears to be a benign product-image audit skill, with a caveat that audience-matching guidance should avoid protected-characteristic targeting.

Before installing, treat the image-audit guidance as a content-quality aid, not a targeting or exclusion tool. If you use it for marketing decisions, add your own rule that recommendations must avoid protected characteristics and use inclusive, non-discriminatory language.

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)

Natural-Language Policy Violations

Medium
Confidence
89% confidence
Finding
The guide explicitly scores lifestyle imagery on whether the model or setting 'matches target audience' without qualifying that protected characteristics must not be used for exclusionary or discriminatory targeting. In a product-page audit skill, this can normalize recommendations based on race, age, gender, or other sensitive traits, creating legal, ethical, and platform-policy risk if users operationalize it in creative decisions.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.