ClawHub
Back to skill

Security audit

创建检查编辑

Security checks for vulnerabilities and agentic risk

Overview

This DOCX skill is review-worthy because it can edit local documents and invoke commands or external APIs without tight scoping or clear overwrite controls.

Install only if you are comfortable giving this skill access to read and modify DOCX files and potentially run local commands. Use copies of important documents, prefer explicit output paths, and avoid processing sensitive files unless you understand which commands and external APIs your agent will use.

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
80% confidence
Finding
The skill gives contradictory trust signals about whether it can execute commands or is only pure Markdown. This is dangerous because users and agent platforms may underestimate system-side effects, leading to unintended command execution, file modification, or unsafe approval decisions.

Missing User Warnings

Medium
Confidence
83% confidence
Finding
The skill is designed to create and edit DOCX files but does not prominently warn that using it may overwrite or modify user documents. This can cause unintended data loss, corruption of important files, or silent modification of sensitive documents when the user assumes analysis-only behavior.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The skill mentions command execution and external API integration without a prominent warning about system and network side effects. This is dangerous because users may provide untrusted paths or content while unaware that local commands may run or data may be sent externally, increasing the risk of data leakage or unintended host actions.

Static analysis

No suspicious patterns detected.