ClawHub
Back to skill

Security audit

网络配置排障

Security checks for vulnerabilities and agentic risk

Overview

This skill is a private-network configuration guide, but it requests command and file-write authority without clear limits or user consent checkpoints.

Review this skill carefully before installing. Use it only when you are prepared to approve each command, file change, network configuration change, and outbound API call; do not let it modify VPN, DNS, credentials, or system network settings automatically.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Intent-Code Divergence

Medium
Confidence
92% confidence
Finding
The skill documentation inconsistently describes itself as both command-executing (`exec`, file/API actions, sandboxed commands) and pure Markdown guidance. This ambiguity can cause an agent or user to authorize higher-risk behaviors than intended, increasing the chance of unintended system changes or command execution during use.

Description-Behavior Mismatch

Medium
Confidence
90% confidence
Finding
The skill is presented as a focused private-network configuration tool, but its documented capabilities expand into generic file handling, API integration, and command execution. This broadens the effective privilege and attack surface of the skill, making it easier for the skill to be repurposed for unintended system actions under a misleadingly narrow description.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill advertises file writing, API calls, and command execution without prominent warnings about modifying the host system, transmitting data externally, or affecting network configuration. In a security-oriented network tool, these capabilities are especially sensitive because they may alter connectivity, expose secrets, overwrite files, or run unsafe commands if invoked by an agent without informed approval.

Static analysis

No suspicious patterns detected.