Back to skill

Security audit

Flight Price Comparison

Security checks for vulnerabilities and agentic risk

Overview

The skill is a flight-price comparison prototype, but it asks users to store travel-site passwords in a general agent file and overstates functionality that is not implemented.

Review before installing. Do not put travel-site passwords in TOOLS.md; use manual or QR login in a dedicated browser profile if you test it. Treat this as an unfinished prototype, not a reliable flight-price comparison tool, until it clearly implements search, price extraction, credential handling, and platform scope.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:125
Finding

Plaintext Travel-Platform Credentials Stored in Agent Configuration

Content
View full analysis
Remediation
View remediation
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 (8)

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

这是明显的描述与实际行为不符。代码的表面目标与声明一致,都是机票比价,但实现程度远低于声明能力。两个平台函数只是打开页面,飞猪函数额外做了页面快照;核心 TODO 明确表明搜索、结果提取和价格比较尚未完成。结果生成函数依赖 lowest_price 和 flights,但搜索函数始终返回空列表和 None,因此无法产出有效比价结果。不存在明显的额外越权或无关恶意能力;问题在于声明把未实现的功能说成已具备。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The file consistently specifies the skill in Chinese and provides no indication that users may interact in other languages or choose a preferred locale. This can violate language/locale policy when a skill forces a specific language without opt-in or documented regional justification.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

L008 写明可自动对比“飞猪、携程、去哪儿等平台”,但 L079-L085 的“支持的平台”表只列出飞猪和携程,清单顶部描述 L003 也只明确提到“飞猪、携程等平台”。这不是单纯信息缺失,因为示例输出 L104-L118 还把去哪儿作为实际比价结果来源,导致文档内部对技能真实支持范围存在冲突。

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The markdown presents activation examples such as “帮我查一下...” and “查一下...” that overlap with ordinary conversational requests, but it does not define what exact phrases or contexts should invoke the skill versus ordinary chat. Without explicit trigger constraints or negative examples, the skill could be invoked unintentionally for general travel discussion.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The top-level documentation claims this is a working 'full-network flight price comparison' script, while the code body contains TODO placeholders and no functional extraction logic. In the context of an agent skill, this mismatch increases risk because the agent may select and invoke the skill under false assumptions about capability, leading to deceptive behavior and unreliable outputs.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This code file contains user-facing natural-language documentation, usage examples, and help text entirely in Chinese. Under the policy, forcing a specific language without offering a user choice or documenting a justified locale constraint is a language/locale policy violation.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill advertises automated cross-platform flight price comparison, but the core search functions do not perform any real search or price extraction and instead return empty results. This is dangerous because downstream agents or users may rely on the tool's claimed output for travel decisions, causing silent failure, misleading recommendations, or incorrect automation behavior.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The statement that credentials 'will not be sent to any third party' is materially misleading because the skill's stated purpose is to log into third-party travel platforms. This can cause users to underestimate credential exposure and trust boundaries, especially since the workflow requires storing sensitive login data in TOOLS.md and then using browser automation against external sites.

Content

No source excerpt is available for this finding.