Back to skill

Security audit

带谁去哪

Security checks for vulnerabilities and agentic risk

Overview

This travel recommendation skill has a coherent goal, but it asks for risky system changes, weakens secure network checks, and stores personal travel profile data beyond what the core task requires.

Review before installing. Do not run the sudo install path, avoid automatic global upgrades from @latest, and do not use the TLS bypass. Only save a travel profile if you are comfortable with family, budget, location, and preference details being reused across sessions or written to ~/.flyai/user-profile.md.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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
Findings (3)

T08 · Insecure Dependencies

Error
Location
SKILL.md:106
Finding

Mandatory Installation of an Unpinned Third-Party CLI Package

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Error
Location
SKILL.md:125
Finding

Privileged Installation of an Unpinned npm Package

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:132
Finding

TLS Certificate Verification Disabled for FlyAI 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
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (26)

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill requires a global npm install/upgrade of a CLI before performing ordinary recommendation tasks, which introduces unnecessary code execution and system modification. Because this is unrelated to the narrow purpose of companion-based attraction recommendations, it expands the attack surface and could expose users to supply-chain compromise or unintended host changes.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

Recommending sudo for a global npm install encourages privileged execution of package installation, which can lead to full-system compromise if the package or install path is malicious or tampered with. For a travel recommendation skill, this privilege escalation is unjustified and materially increases risk.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
100% confidence
Finding

The skill explicitly instructs disabling TLS certificate validation with NODE_TLS_REJECT_UNAUTHORIZED=0. This defeats transport security, making requests vulnerable to man-in-the-middle interception or tampering and exposing query contents and returned data integrity.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The instructions recommend disabling TLS verification without explaining the confidentiality and integrity risks. This normalizes an unsafe workaround that can expose travel queries, account context, and search results to interception or manipulation.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The exception-handling table again advises bypassing SSL verification, this time as routine troubleshooting, without any disclosure of the serious privacy and integrity consequences. Repetition in fallback guidance makes unsafe behavior more likely to be adopted in practice.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The activation text says to use the skill when the user mentions phrases like "适合XX的景点", "同行人", or "家庭出游", which are broad and may appear in many general travel queries. The description does not provide exclusion conditions or clearer scope boundaries for when this skill should activate instead of a more generic travel skill.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs shell execution and global package installation without prominently warning users that it will modify the system. Lack of disclosure undermines informed consent and increases the chance that users or agents perform risky host changes unexpectedly.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
97% confidence
Finding

The explicit use of sudo is a least-privilege violation. Running npm installation as root grants extensive system access to install scripts and dependencies, magnifying the impact of any malicious or compromised package behavior.

Content

Scanner excerpt · SKILL.md (reported line 125)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` |
| 网络问题 | 建议用户检查网络,确保可以访问 registry.npmjs.org |

### Step 1: 搜索景点

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
100% confidence
Finding

Setting NODE_TLS_REJECT_UNAUTHORIZED=0 creates an unsafe default that disables certificate validation for the process. This directly weakens transport security and can allow adversaries to spoof the remote endpoint or alter returned POI data.

Content

Scanner excerpt · SKILL.md (reported line 134)May include surrounding context.

bash
# 注意:如遇SSL证书验证失败,添加环境变量绕过
NODE_TLS_REJECT_UNAUTHORIZED=0 flyai search-poi \
  --city-name "[目的地]"

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
100% confidence
Finding

Repeating NODE_TLS_REJECT_UNAUTHORIZED=0 as standard error handling entrenches an insecure operational default. In the context of a non-security-critical travel skill, there is no justification for trading away transport authenticity and privacy to keep the flow running.

Content

Scanner excerpt · SKILL.md (reported line 233)May include surrounding context.

md
| POI无适龄标签 | 基于景点类型推理(博物馆=室内低体力、山岳=爬山高消耗) |
| 目的地景点太少 | 扩大搜索范围到周边城市,或使用 keyword-search 补充 |
| 用户坚持去不推荐的景点 | 给出注意事项和降低难度的方式 |
| SSL证书验证失败 | 使用 `NODE_TLS_REJECT_UNAUTHORIZED=0` 环境变量绕过 |

## 工具说明

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The manifest describes a skill focused on screening destinations and attractions based on companion characteristics, with FlyAI attraction data and related hotel search as the core. The documented follow-up actions add flight and train search capabilities, which materially broaden the skill from companion-based attraction recommendation into transportation planning.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The referenced documentation exposes a broad semantic travel search capability that goes beyond the skill's stated purpose of companion-based attraction filtering. This mismatch can cause the agent to invoke a much more general tool than intended, increasing the risk of scope creep, unintended data retrieval, and recommendations outside the skill's guardrails.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This markdown file contains user-facing skill content exclusively in Chinese, including headings, prompts, and outputs. Under the policy rule for natural-language violations, forcing a specific language without user opt-in can be a locale policy issue when no alternative language option or justification is provided.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The manifest says this skill recommends destination attractions based on companion profiles such as children, elders, or friends, and filters scenic spots accordingly. This file instead documents a hotel-search command with lodging-specific parameters like check-in/check-out dates, star ratings, bed types, and nightly price, which is a materially different user-facing capability.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The entire tool documentation, including all user-facing example questions, is written in Chinese and presents the interaction model as Chinese-only. There is no indication that users may choose another language or that the locale restriction is intentional for a region-specific use case, which violates the language/locale policy criterion.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The document defines cross-session and effectively cross-skill user profile persistence for travel preferences, even though this companion-matching skill is primarily a recommendation/filtering helper. Persisting profile data broadens the skill's authority and data exposure surface beyond what is necessary, creating privacy and misuse risk if other skills, agents, or local processes can access the stored data.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The specification describes storing and updating user travel profile data across sessions but does not disclose the privacy implications of persistence, reuse, or possible access by other components. Users may unknowingly share long-lived personal preference and family information, preventing informed consent and increasing privacy harm.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The local fallback instructs storing personal travel profile information in ~/.flyai/user-profile.md, which creates durable local storage of personal data unrelated to the minimum needed to answer a single recommendation request. This increases exposure to other local users, malware, backup/sync systems, or unrelated tooling that can read files in the user's home directory.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Writing profile data to ~/.flyai/user-profile.md without warning about local disk persistence exposes personal information such as city, budget, travel history, family composition, and special needs to local processes and anyone with filesystem access. The absence of notice and safeguards makes accidental overcollection and silent retention more likely.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
88% confidence
Finding

The instruction to create ~/.flyai and persist a user-profile file establishes session persistence of personal data, allowing information provided in one interaction to affect future interactions without robust consent or access controls. In the context of a recommendation skill, this persistence is more dangerous because it is not strictly required for core functionality and includes potentially sensitive household details.

Content

Scanner excerpt · reference/user-profile-storage.md (reported line 73)May include surrounding context.

如果 ~/.flyai/ 目录不存在,需要先创建:

bash
mkdir -p ~/.flyai

文件格式

Vague Triggers

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

The documentation describes capabilities in broad terms without clear activation boundaries or exclusion cases, which can make the agent over-trigger this skill for loosely related travel requests. In an agent setting, ambiguous scope can lead to inappropriate tool use and lower-assurance behavior even if it is not directly exploitable as code execution.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The only natural-language examples shown are in Chinese, which can imply that the skill expects or prefers a specific language. There is no accompanying note that other languages are supported or that Chinese is merely illustrative, so this may violate language-choice expectations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

This markdown file contains user-facing examples and output fields using Chinese city names and Chinese-language values such as weekday and journey type labels. Because the file does not state that the skill is region-specific or offer an alternate language/locale, it may violate the policy against forcing a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The file presents required/expected user inputs and examples using Chinese terms such as bed-type values and destination/hotel names, which effectively constrains usage to a specific language/locale. There is no explicit opt-in, alternative language guidance, or justification that this skill is intended only for a Chinese-market workflow.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The examples and sample output are entirely in Chinese, including search keywords and returned hotel/package fields, with no indication that users may choose another language or locale. This can violate language-choice policy when a skill implicitly assumes a specific language without opt-in or documented regional limitation.

Content

No source excerpt is available for this finding.