ClawHub
Back to skill

Security audit

日历管理工具包基础版

Security checks for vulnerabilities and agentic risk

Overview

The skill is a calendar assistant, but it gives broad execution and data-flow guidance without clearly scoping when calendar data, credentials, callbacks, or external APIs may be used.

Review this before installing if your calendar contains sensitive meetings or contacts. Use it only with explicit calendar tasks, avoid callback URLs unless you trust the destination, and confirm what provider credentials, local cache, and cloud sync data the agent will access before allowing it to run commands or API calls.

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 (4)

Description-Behavior Mismatch

Medium
Confidence
92% confidence
Finding
The skill states that FREE-version data is stored locally and not uploaded to the cloud, but elsewhere it explicitly supports external API usage and cross-platform sync. This contradiction can mislead users and operators about where sensitive calendar data may flow, creating privacy and trust risks if the agent transmits event details, participant data, or tokens to third-party services.

Vague Triggers

High
Confidence
90% confidence
Finding
The trigger condition says to use the skill for data analysis, reporting, statistics, and visualization, which is substantially broader than a calendar-management skill. In an agent environment with exec and network-capable tooling, such overbroad activation criteria can cause unintended invocation on unrelated tasks, increasing the chance of unnecessary command execution, external API calls, or privacy-impacting handling of non-calendar data.

Vague Triggers

Medium
Confidence
83% confidence
Finding
The natural-language invocation guidance is extremely generic and tells the agent to execute corresponding operations based on user input without strong scope boundaries. In a skill that advertises exec capability and API integrations, vague invocation rules raise the risk that unrelated or unsafe requests will be interpreted as permission to perform commands or transmit data externally.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill accepts a callback_url and describes automatic execution of tools or APIs, but it does not clearly warn users about network access, external API transmission, command execution, or callback-driven data flows. This lack of disclosure is dangerous because users may provide sensitive scheduling data or credentials without understanding that the agent could send them to remote systems or invoke local commands.

Static analysis

No suspicious patterns detected.