ClawHub
Back to skill

Security audit

立白龙虾购

Security checks for vulnerabilities and agentic risk

Overview

This shopping skill is mostly coherent, but it runs sensitive login, address, order, and payment-related actions through an unpinned external npm command with unclear provenance and broad activation triggers.

Install only if you trust the publisher and can verify `@libydic/mall` is the legitimate Liby mall tool. Use it intentionally for mall tasks, confirm products, prices, points/cash requirements, addresses, and orders before payment, and avoid sharing phone codes or address details unless the login flow and package provenance are clear.

SkillSpector

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 (2)

Intent-Code Divergence

Medium
Confidence
87% confidence
Finding
The skill says login must only use the auth-start QR flow, but it also documents a direct login step using mobile number and verification code. This inconsistency can cause the agent to collect sensitive credentials in chat and weakens the intended authentication boundary, increasing phishing and credential-handling risk.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger list contains broad everyday terms that can activate the shopping skill in unrelated conversations. Because this skill can access account, order, address, and purchase flows, accidental invocation may expose personal commerce context or steer users into unintended transactional actions.

Static analysis

No suspicious patterns detected.