ClawHub
Back to skill

Security audit

163邮箱管理专家

Security checks for vulnerabilities and agentic risk

Overview

This skill is mostly a disclosed 163.com email-management guide, but it gives an agent broad email send/read/delete authority with weak scoping around destructive and generic messaging actions.

Review before installing. Use this only for a 163.com mailbox you control, avoid configuring unrelated API keys, store the mailbox authorization code carefully, and require explicit confirmation before any send, move, attachment download, or bulk delete operation.

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 (4)

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The document classifies the skill as pure Markdown/natural-language driven while earlier sections explicitly instruct the agent to execute CLI commands. This mismatch can bypass platform safety assumptions, causing an agent or reviewer to grant the skill broader trust than warranted for an EXEC-capable integration.

Intent-Code Divergence

Medium
Confidence
89% confidence
Finding
The dependency section claims an LLM API key is required even though the described functionality is an IMAP/SMTP CLI mail tool. This inconsistency can mislead users into exposing unnecessary secrets or configuring unrelated credentials, increasing the attack surface and the chance of credential mishandling.

Vague Triggers

Medium
Confidence
86% confidence
Finding
The invocation text is broad enough to match generic messaging, notification, and communication tasks beyond 163 mailbox administration. Overbroad routing increases the chance the agent will invoke this skill in unintended contexts, leading to unauthorized email actions, data exposure, or misuse for bulk messaging.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill provides bulk deletion instructions for mail folders, including deleting all messages, without a strong irreversible-deletion warning or confirmation requirement. In an agent context, such guidance can normalize destructive operations and increase the risk of accidental or unauthorized mailbox data loss.

Static analysis

No suspicious patterns detected.