ClawHub
Back to skill

Security audit

litigation response

Security checks across malware telemetry and agentic risk

Overview

This markdown-only litigation-response skill is coherent and purpose-aligned, with the main caution being sensitive legal and personal data handling.

Before installing or using this skill, treat lawsuit documents as sensitive. Redact unnecessary ID numbers, addresses, contact details, bank data, and confidential facts where possible, and verify any legal strategy with a qualified lawyer.

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
91% confidence
Finding
The skill explicitly instructs extraction and structuring of highly sensitive personal and litigation data such as ID numbers, addresses, contact details, and case materials, but provides no privacy warning, minimization guidance, or redaction requirement. In a legal-analysis context, this increases the risk of unnecessary disclosure of personally identifiable information and confidential case details to the agent or downstream systems.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.