ClawHub

US Car Insurance Advisor

Security checks across malware telemetry and agentic risk

Overview

This is a US-focused car insurance advice skill made of reference documents, with no code, install scripts, credential use, or hidden data access.

Use this as general US auto-insurance guidance, not a substitute for your policy documents, insurer, attorney, or licensed insurance professional. Verify state-specific rules and claim deadlines before acting, and avoid sharing exact personal identifiers unless needed.

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 (2)

Natural-Language Policy Violations

Medium
Confidence
96% confidence
Finding
The section is labeled as applying to 'All Claim Types' and presents emergency, police-reporting, and insurer-contact steps as universal, but these are jurisdiction-specific and assume US norms. In a safety- and claims-related skill, users may follow this as authoritative guidance outside the intended region, leading to legal noncompliance, delayed claims, or unsafe handling of emergencies where local procedures differ.

Natural-Language Policy Violations

Medium
Confidence
93% confidence
Finding
The later sections use dense US-specific legal and insurance concepts such as PIP, UM/UIM, no-fault states, rideshare coverage phases, Department of Insurance complaints, and state windshield laws without reinforcing geographic scope. Because the skill provides actionable claims advice, a user outside the US or in a different state could misapply the guidance and make harmful financial or legal decisions based on inapplicable rules.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal