ClawHub
Back to skill

Security audit

网页

Security checks for vulnerabilities and agentic risk

Overview

This web-development skill is broad but coherent, with expected file, command, and deployment capabilities and no evidence of hidden execution or data theft.

Before installing, treat this as a broad web-development helper that may write project files, run commands, use API keys, and assist with deployment. Use it only in intended project directories, review commands before execution, and do not assume the stated sandbox or whitelist exists unless your agent platform enforces it.

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 (2)

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The skill claims command execution is limited by a whitelist or safe sandbox, yet the declared tool set includes unrestricted exec and the document provides no enforceable mechanism for those restrictions. This can mislead users or downstream agents into trusting shell execution that may actually run arbitrary commands, increasing the chance of filesystem, credential, or environment compromise.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The skill advertises deployment, file writing, command execution, and external API usage without a prominent warning that these actions can modify local systems, expose secrets, or affect live infrastructure. In an agent setting, understated side effects can cause unsafe user consent and increase the likelihood of unintended destructive or privacy-impacting actions.

Static analysis

No suspicious patterns detected.