Back to skill

Security audit

机票低价日历

Security checks for vulnerabilities and agentic risk

Overview

The skill is a useful flight-price calendar, but it tells agents to bypass TLS security and install an unpinned global CLI, which creates review-worthy risk.

Review carefully before installing. Do not run the provided commands with NODE_TLS_REJECT_UNAUTHORIZED=0, avoid sudo/global @latest npm installation, and treat returned booking links as untrusted unless the CLI and destination domains are independently verified.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (2)

T08 · Insecure Dependencies

Warning
Location
reference/core-workflow.md:3
Finding

Automatic Installation of an Unpinned Global Dependency

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
reference/core-workflow.md:41
Finding

TLS Certificate Verification Disabled for Travel Search Requests

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (35)

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The instruction to use NODE_TLS_REJECT_UNAUTHORIZED=0 disables TLS certificate verification, allowing man-in-the-middle attacks against external API connections. In a flight-booking context this could expose itinerary details, search parameters, tokens, or manipulated pricing data, and it is unrelated to the legitimate function of a price calendar assistant.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The referenced documentation describes a general-purpose AI travel semantic search tool rather than the declared flight low-price calendar skill. This mismatch can cause the agent to invoke a capability broader than intended, increasing the chance of over-collection of user data, unintended actions outside scope, or misleading users about what the skill actually does.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The workflow repeatedly instructs operators to set NODE_TLS_REJECT_UNAUTHORIZED=0 for flight and related searches, which disables TLS certificate validation for network traffic. This enables man-in-the-middle interception or tampering of API responses, including prices, booking links, and travel data, and is especially risky because it is presented as a normal required step rather than a narrow temporary workaround.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The markdown explicitly tells users to bypass SSL certificate verification without warning about the security consequences. Presenting insecure transport as standard operating procedure normalizes unsafe behavior and materially increases the chance that users will expose booking sessions and returned data to interception or manipulation.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The referenced file documents a hotel search capability even though the skill is declared as a flight low-price calendar assistant. This kind of scope mismatch is dangerous because it can expose unintended tool capabilities, broaden the agent’s action surface, and enable invocation of unrelated travel functions that users did not consent to or expect.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The referenced documentation is for Marriott hotel search, while the declared skill is a flight low-price calendar assistant. This mismatch can cause the agent or integrator to invoke the wrong capability, route user travel data to an unintended function, or produce misleading behavior that bypasses expected safety and business constraints. In a travel-booking context, capability confusion is especially risky because users may disclose itinerary and location data under false assumptions.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The reference file documents a POI/attraction search capability, while the skill is described as a flight low-price calendar assistant. This kind of tool/skill mismatch can cause the agent to invoke the wrong capability, mis-handle user intent, or expose unrelated data flows and parameters, which is especially risky because it changes the operational scope of the skill without notice.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The reference file documents a train-search command for a skill whose stated purpose is a flight low-price calendar. This inconsistency can cause the agent to invoke the wrong capability, mishandle user requests, or expose unintended tool behavior, which is a real security and safety issue in agent systems because tool-selection confusion can lead to unauthorized or misleading actions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The description and trigger phrases are entirely in Chinese and the prescribed user interaction text is also fixed in Chinese. There is no indication that the skill offers language choice or that the Chinese-only behavior is a documented, justified regional constraint.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill reads and later persists a user travel profile at startup even though its primary purpose is flight-price scanning and calendar generation. This creates unnecessary access to personal preference data and expands data collection beyond what is required for the task, increasing privacy and data-handling risk if the profile contains sensitive or identifying information.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill accesses user profile data automatically without clearly warning users in the visible skill description that their stored preferences may be read. Lack of transparent notice and consent can lead to privacy violations and unexpected processing of personal data, especially because the fallback path reads a local profile file.

Content

No source excerpt is available for this finding.

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
99% confidence
Finding

Setting NODE_TLS_REJECT_UNAUTHORIZED=0 is an unsafe default because it globally weakens transport security for HTTPS requests. This makes the skill substantially more dangerous in context: it interacts with external flight data and booking-related services, so compromised TLS can lead to data exposure, response tampering, and user deception through manipulated results.

Content

Scanner excerpt · SKILL.md (reported line 106)May include surrounding context.

md
| 去程便宜但返程贵 | 计算往返组合价格,避免只看单程误导 |
| 全部日期都很贵 | 提示"该时段为旺季,建议:① 提前2个月关注 ② 考虑错峰 ③ 尝试邻近机场" |
| API 调用失败 | 使用历史同期估价,标注"参考价" |
| SSL 证书验证失败 | 确保使用 `NODE_TLS_REJECT_UNAUTHORIZED=0` 前缀 |

---

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The entire skill file is written in Chinese and provides no indication that users may choose another language or that the skill is intentionally limited to a Chinese-speaking context. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The entire workflow mandates Chinese-language headings, labels, and output templates, but does not state that the skill is region-specific or provide any language opt-in. Under the policy, forcing a specific language without user choice is a natural-language locale violation unless clearly justified.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
75% confidence
Finding

The documentation recommends using sudo npm install -g to install a package from a public registry. Running a global package installation with elevated privileges increases the blast radius of any malicious package, dependency compromise, install script abuse, or typo-squatting event, potentially leading to full system compromise.

Content

Scanner excerpt · reference/core-workflow.md (reported line 20)May include surrounding context.

md
| 情况 | 处理方式 |
|-----|---------|
| npm 未安装 | 提示用户先安装 Node.js (https://nodejs.org/) |
| 权限不足 | 建议使用 `sudo npm install -g @fly-ai/flyai-cli@latest --registry=https://registry.npmjs.org` 或使用 nvm 管理 Node |
| 网络问题 | 建议用户检查网络或使用国内镜像 `npm config set registry https://registry.npmmirror.com` |

**注意:** 此步骤只在首次使用时执行,后续调用会直接跳过已安装的情况。

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
99% confidence
Finding

This command embeds NODE_TLS_REJECT_UNAUTHORIZED=0 as the default way to run a networked search. Disabling certificate verification breaks transport authenticity and allows attackers on the network path to spoof endpoints or alter flight-search responses.

Content

Scanner excerpt · reference/core-workflow.md (reported line 44)May include surrounding context.

单日查询命令:

bash
NODE_TLS_REJECT_UNAUTHORIZED=0 flyai search-flight \
  --origin "{出发城市}" \
  --destination "{目的地}" \
  --dep-date {出发日期} \

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
99% confidence
Finding

The workflow codifies an unsafe default by telling users they need to prefix commands with NODE_TLS_REJECT_UNAUTHORIZED=0. Because this is framed as required behavior, it is likely to be copied broadly, turning a transient certificate problem into a persistent insecure operating mode.

Content

Scanner excerpt · reference/core-workflow.md (reported line 65)May include surrounding context.

md
| `--dep-hour-start/end` | 出发时间范围 (0-23) |
| `--max-price` | 价格上限 |

> **⚠️ SSL 证书问题**:执行 flyai CLI 命令时,需要在命令前加上 `NODE_TLS_REJECT_UNAUTHORIZED=0` 来绕过 SSL 证书验证。

**批量扫描策略**:

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest scopes the skill to finding the cheapest flight departure dates and generating a low-price flight calendar. Step 5 adds follow-on capabilities for search-hotel and keyword-search for flight+hotel packages, which goes beyond the described flight-calendar assistant behavior.

Content

No source excerpt is available for this finding.

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
98% confidence
Finding

The insecure TLS-bypass default is repeated in a follow-on flight-search command, extending the exposure to subsequent user actions. Repetition across workflow steps increases the chance of habitual insecure use and broadens the window for traffic interception and response tampering.

Content

Scanner excerpt · reference/core-workflow.md (reported line 145)May include surrounding context.

  1. 查看详细航班信息:

    bash
    NODE_TLS_REJECT_UNAUTHORIZED=0 flyai search-flight --origin "{出发城市}" --destination "{目的地}" --dep-date {推荐日期} --back-date {返程日期}
    
  2. 搜索酒店价格:

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
98% confidence
Finding

The same TLS-verification bypass is applied to hotel search, meaning additional travel-related queries and potentially more user data are sent without authenticating the remote endpoint. This widens the affected attack surface beyond the original flight-search use case.

Content

Scanner excerpt · reference/core-workflow.md (reported line 150)May include surrounding context.

  1. 搜索酒店价格:

    bash
    NODE_TLS_REJECT_UNAUTHORIZED=0 flyai search-hotel --dest-name "{目的地}" --check-in-date {入住日期} --check-out-date {退房日期}
    
  2. 搜索机+酒套餐:

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
98% confidence
Finding

The package-search command also disables TLS verification, propagating the same insecure default across another networked feature. An attacker could tamper with search results, links, or content returned by the command when the endpoint identity is not being verified.

Content

Scanner excerpt · reference/core-workflow.md (reported line 155)May include surrounding context.

  1. 搜索机+酒套餐:
    bash
    NODE_TLS_REJECT_UNAUTHORIZED=0 flyai keyword-search --query "{目的地} 机票酒店套餐 {日期}"
    

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This markdown example is entirely written as a Chinese-only interaction and output format, with no indication that the user can choose another language or locale. Under the policy rule, forcing a specific language without user opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The reference file describes a generic travel keyword-search capability spanning hotels, visas, tours, dining, cruises, and telecom rather than a narrowly scoped flight low-price calendar function. This kind of scope drift can cause the agent to invoke unrelated search behaviors on broad user input, increasing the chance of unintended tool use, over-collection of user intent, and responses outside the skill’s declared purpose.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The parameter guidance explicitly allows broad natural-language queries and many open-ended categories, which weakens input boundaries for the skill. In the context of a flight-calendar assistant, overly permissive activation criteria can trigger irrelevant searches or let ambiguous prompts be interpreted too broadly, reducing predictability and increasing the risk of unauthorized or unintended capability expansion.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

Hotel search capability is not justified by the skill’s stated purpose of finding the cheapest flight dates. Even if only documented, this can mislead orchestration logic or maintainers into enabling unrelated operations, increasing the chance of unauthorized data access, confusing behavior, or privilege creep within the travel domain.

Content

No source excerpt is available for this finding.