ClawHub
Back to skill

Security audit

网页抓取工具免费版

Security checks across malware telemetry and agentic risk

Overview

This web-scraping skill is mostly purpose-aligned, but its trigger wording and file-writing behavior are too broad and underspecified for automatic installation.

Review before installing. Use it only for explicit, user-directed scraping of a single public URL, choose output paths carefully, avoid saving sensitive or authenticated page content, and consider narrowing the trigger text and file-save policy before use.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Intent-Code Divergence

Low
Confidence
82% confidence
Finding
The trigger condition is unrelated to the skill's actual web-scraping purpose and could cause the agent to invoke this skill for broad analytics or reporting requests. In an agent environment with exec, read, write, and network access, overly broad invocation criteria increase the risk of unintended scraping, unnecessary file creation, and execution in the wrong user context.

Vague Triggers

High
Confidence
92% confidence
Finding
An overly broad and mismatched trigger condition is a real security quality issue because it can cause unintended invocation of a skill that has exec, write, and network capabilities. In practice, this expands the attack surface for prompt-driven tool misuse: unrelated requests may be routed into a workflow that fetches remote content and writes files, which is more dangerous in this context than in a read-only skill.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The skill advertises saving scraped results to files but does not clearly describe file-writing behavior, path constraints, or overwrite semantics. In a tool-enabled agent, unclear write behavior can lead to accidental overwrites, writing sensitive scraped content to unsafe locations, or unexpected persistence of remote data on disk.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.