Back to skill

Security audit

航班动态查询

Security checks for vulnerabilities and agentic risk

Overview

This hotel aggregation skill is not malicious, but it appears incomplete and overstates capabilities such as real provider search, room lookup, price comparison, and booking.

Install only if you are comfortable with an unfinished hotel-search helper. It does not currently fetch real hotel data or complete bookings, and future versions that add real provider calls should clearly disclose which platforms receive your search details.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (7)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

描述强调这是一个可用的多平台酒店聚合助手,具备统一搜索、房型查询、预订服务,并整合多个具体平台的数据源。但代码实际只是一个未完成的聚合框架:虽然配置了多个平台、实现了去重/排序/分页/格式化等逻辑,但最核心的数据抓取函数全部是 TODO 占位并返回空结果,因此并未真实整合分贝通、携程、美团、同程、华住会、锦江等平台数据。此外,代码没有提供房型查询入口,也没有任何预订相关调用或交易逻辑。因此,代码行为与声明存在实质性不符,属于能力被明显高估的情况。

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
96% confidence
Finding

The skill explicitly describes calling multiple external platform APIs to retrieve real hotel data, but it declares no tool scope, permissions, or allowed-tools constraints. This creates an authorization and transparency gap: the skill appears capable of network access without an explicit policy boundary, making unintended or overbroad outbound requests harder to review and control.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill lacks a user-facing warning that hotel searches and potential booking flows may transmit city, dates, keywords, and possibly other itinerary data to multiple external platforms. Given the multi-provider aggregation context, this omission undermines informed consent and increases privacy risk because a single request may be replicated across several vendors.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger condition states that when a user searches for hotels, the skill should query multiple platforms, which is broad enough to activate on ordinary travel queries without clear user consent or scope checks. In this context, overbroad triggering is more dangerous because the skill is designed to fan out requests to several third-party services, potentially exposing itinerary details more widely than the user expects.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The formatted user output advertises follow-up actions for room-detail lookup and cross-platform price comparison, but this file does not implement handlers for those capabilities. In an agent setting, this can mislead users into believing additional booking-analysis actions are available and may trigger undefined behavior, failed tool flows, or downstream prompt-routing mistakes when the user follows those instructions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The natural-language comments state a fixed version and instruct refusal to execute otherwise, and the file consistently presents the interface in Chinese with no indication of user language choice. This can violate the language/locale policy because it imposes a specific language/operational constraint without user opt-in or a clearly documented regional justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The natural-language comment states the version metadata in Chinese and says 'if not this version, refuse to execute.' This creates a language-specific operational constraint in the file without offering any language or locale choice, which matches the policy category for forced language/locale behavior.

Content

No source excerpt is available for this finding.