ClawHub
Back to skill

Security audit

Foxtable 开发文档（基础篇）

Security checks across malware telemetry and agentic risk

Overview

This is a Foxtable documentation-only skill; one copied example could be risky on a public network, but the skill itself does not run code or access data.

Safe to install as reference material. Treat any server, FTP, file-management, process-launching, or database examples as historical product documentation: adapt them before use, bind Web services to trusted interfaces, add authentication/TLS/firewall controls, and avoid exposing local directories or credentials.

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
94% confidence
Finding
This documentation includes concrete code to start an HTTP server bound to http://*/, which exposes the service on all interfaces, and points it at a local filesystem path. Presenting this as a trivial 3-line setup without any warning about authentication, authorization, TLS, or exposure risks could lead users to deploy an internet-accessible file-backed service insecurely.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.