ClawHub
Back to skill

Security audit

文件上传，本地文件转网络路径

Security checks for vulnerabilities and agentic risk

Overview

This skill does upload files as advertised, but it under-explains where files go, how public the returned links are, and how much data it may upload.

Review before installing. Use it only for non-sensitive files you are comfortable uploading to the actual service endpoint, confirm who operates that endpoint, whether links are public or expiring, and how to delete uploaded files. Avoid using broad or long-lived JWT tokens unless you know how to revoke them.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Intent-Code Divergence

Medium
Confidence
97% confidence
Finding
The skill documentation states a 4MB upload limit, but the script actually allows files up to 100MB. This mismatch is security-relevant because users and reviewers may rely on the documented constraint while the implementation permits substantially larger uploads, increasing bandwidth, storage, and abuse exposure.

Vague Triggers

Medium
Confidence
82% confidence
Finding
The trigger phrases are broad and overlap with ordinary user requests such as uploading images or documents. In the context of a skill that transfers local files to a remote service and returns public URLs, overbroad activation can cause unintended data disclosure or accidental publication of sensitive files.

Missing User Warnings

High
Confidence
95% confidence
Finding
The skill is designed to upload local files to a remote OSS-backed service and produce directly accessible URLs, but the documentation does not warn users about privacy, confidentiality, or permanence risks. In this context, missing disclosure is dangerous because users may unintentionally publish personal, proprietary, or regulated data.

Static analysis

No suspicious patterns detected.