ClawHub
Back to skill

Security audit

geo-report

Security checks across malware telemetry and agentic risk

Overview

This skill is a client-report generator that reads GEO audit outputs and writes a consolidated report; its main risk is that business-impact estimates may need human review before client delivery.

Before sending generated reports to clients, review any traffic, citation, or revenue projections and ensure they are clearly tied to first-party analytics or stated assumptions. The skill appears appropriate for report generation, but business-impact claims should not be treated as verified financial forecasts without human review.

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
96% confidence
Finding
The skill repeatedly directs the agent to produce client-facing traffic, citation, and revenue uplift estimates as authoritative output, while only lightly saying to be conservative and never guarantee results. That creates a real risk of deceptive or unsupported business claims being presented as audit conclusions, which can mislead customers and drive poor commercial decisions or liability for the operator.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.