ClawHub
Back to skill

Security audit

Web Scraper

Security checks across malware telemetry and agentic risk

Overview

This is mostly a coherent web-scraping skill, but it needs review because it instructs agents to reveal soft-paywalled content hidden in the page.

Install only if you will use it on sites and content you are authorized to scrape. Review generated scripts before running them, avoid the soft-paywall reveal workflow, and do not use optional OpenRouter entity extraction on confidential or access-controlled content unless that external data flow is acceptable.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Intent-Code Divergence

High
Confidence
94% confidence
Finding
The document claims the skill never makes direct API calls itself, yet later includes code that performs a direct POST to OpenRouter. This inconsistency can mislead reviewers and operators about data egress, trust boundaries, and credential use, increasing the chance that sensitive scraped content is sent to a third party unexpectedly.

Intent-Code Divergence

Medium
Confidence
97% confidence
Finding
The skill says it will not bypass paywalls, but then instructs the agent to remove soft-paywall overlays and reveal hidden content already present in the DOM. That is effectively guidance to circumvent an access-control mechanism for subscriber-restricted material, which can enable unauthorized extraction of protected content.

Context-Inappropriate Capability

High
Confidence
98% confidence
Finding
The skill explicitly enables circumvention of soft paywalls, which goes beyond ordinary web scraping and enters access-control bypass behavior. Even if the content is present client-side, using automation to reveal hidden subscriber content undermines publisher restrictions and could create legal, policy, and abuse risks.

Ssd 4

Medium
Confidence
96% confidence
Finding
The paywall section cumulatively steers the agent toward using DOM manipulation to access subscriber-restricted material that is intentionally obscured by the site. That materially increases abuse potential by operationalizing a method to defeat content gating, even if no authentication is broken.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.