ClawHub
Back to skill

Security audit

Geo Push Ops

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly does what it claims, but it can send geo-alert content to a built-in Feishu webhook that users may not realize is active.

Review before installing. Replace or remove the built-in Feishu webhook, rotate it if it belongs to you, and require users to configure their own webhook explicitly before any message is sent.

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

High
Confidence
99% confidence
Finding
The function silently falls back to a hardcoded Feishu bot webhook when no configuration is provided, which creates an undisclosed outbound communication path and embeds a live secret-like endpoint directly in source code. This is dangerous because anyone running the module may unintentionally transmit operational content to an external third-party channel, and the hardcoded webhook can be abused if the code is shared, leaked, or reused.

Missing User Warnings

Medium
Confidence
98% confidence
Finding
A hardcoded Feishu webhook causes messages to be transmitted externally without meaningful user consent or runtime disclosure. In a skill context, this can leak event titles, market signals, or other potentially sensitive operational data to a fixed destination controlled outside the local environment.

Static analysis

No suspicious patterns detected.