Back to skill

Security audit

帮我凑假

Security checks for vulnerabilities and agentic risk

Overview

This travel-planning skill is mostly coherent, but it asks agents to install an unpinned global CLI and read or store personal travel-profile data, which needs careful review before use.

Install only if you trust the FlyAI CLI publisher and are comfortable with a global npm package being installed or upgraded on your machine. Do not run the install with sudo unless you fully accept system-wide risk. Treat saved travel profiles as personal data, especially budget, family, travel history, and accessibility needs, and prefer one-time use if you do not want that data persisted.

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

Error
Location
SKILL.md:27
Finding

Mandatory Installation of an Unpinned Global Dependency

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:101
Finding

Potential Shell Command Injection Through Interpolated User Input

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • 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
Findings (21)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The activation phrases are broad enough to overlap with ordinary travel-planning queries, which can trigger the skill in situations beyond its narrowly described use. Overbroad invocation increases the likelihood of unnecessary tool usage, data reads, or side effects such as profile access or install instructions being surfaced when the user did not intend to use this skill.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest presents the skill as a leave-planning and flight-price assistant, but the body expands scope into hotels, POI lookup, destination guides, visa checks, and package search. This mismatch can cause users or orchestrators to grant the skill broader trust and permissions than expected, increasing the chance of unintended tool use and data access.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill directs the agent to install or upgrade a global npm package before performing normal travel-planning tasks, which is an unnecessary system modification for the stated purpose. This expands impact from simple content generation to arbitrary code execution from an external package source, creating supply-chain and host-compromise risk if the package or dependency chain is malicious or tampered with.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill instructs a global CLI installation/upgrade without prominently warning that it modifies the system, adds software, and may execute third-party package scripts. In the context of an agent skill, this is especially dangerous because users expect planning assistance, not host-level changes, and may not realize the security implications of allowing the step.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill reads persisted user profile data from memory or a local file and later suggests saving new preferences, even though a one-shot leave-planning interaction does not require broad access to stored personal data. This creates unnecessary privacy exposure and increases the chance of over-collection, unintended retention, or leakage of travel preferences and related personal information.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The skill accesses persisted user profile data without an explicit privacy disclosure in the user-facing description, so users may not understand that their local files or stored memory will be consulted. Hidden data access is risky in a travel context because preferences, locations, and trip history can be sensitive and may be processed without meaningful consent.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest says this skill is for generating optimal leave-taking plans and querying flight prices based on destination, leave days, and departure city. This reference file instead documents a generic 'ai-search' capability covering hotels, attractions, flights, trains, mixed travel intent, budgets, companions, and itinerary recommendations, which is materially broader than leave-planning.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The examples and allowed category values are presented as Chinese-only inputs such as 西湖, 故宫, and category enums like 自然风光, which implies the skill expects a fixed locale/language. The file does not mention any user choice, alternative locale support, or justification for restricting input language, which can violate language/locale policy requirements.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The document adds persistent collection and storage of user travel-profile data across sessions, including fallback to a local file, which exceeds the skill’s stated purpose of generating leave plans and querying flight prices. This creates unnecessary data retention and broadens exposure of personal data without clear necessity, increasing privacy and misuse risk if the host environment or filesystem is accessed by others.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The storage instructions tell the agent to read and update persistent user-profile data but do not lead with a privacy notice, consent requirement, or retention policy. As a result, users may have their personal preferences and family-related information stored without understanding where it is kept, for how long, or who can access it.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
89% confidence
Finding

The instruction to create ~/.flyai and persist a user profile in ~/.flyai/user-profile.md establishes session persistence on the local filesystem, potentially storing personal data in plaintext. On shared or insufficiently secured systems, this can expose travel preferences, family details, and special requirements to other local users, backup systems, or unrelated tooling.

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

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The spec directs storage of detailed personal attributes such as family composition, child age, budget, travel history, and special needs, which are sensitive or highly profiling in nature and not clearly justified for a vacation-planning assistant. Persisting this information in memory or a local markdown file increases the chance of privacy harm, profiling abuse, or unintended disclosure across tools, sessions, or users on the same machine.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The only user-facing query examples are in Chinese, which implicitly steers usage toward a specific language/locale. The file does not state that other languages are supported, offer a language choice, or justify a China-specific scope, so this can be read as a locale preference without opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

This markdown example is entirely presented in Chinese and labels the assistant response as the expected output, which can imply a fixed language behavior. The policy requires avoiding language/locale constraints unless the user is given a choice or the constraint is clearly documented and justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

This markdown file presents all user-facing content in Chinese and does not indicate that language selection is optional or that the skill is intentionally limited to a Chinese-speaking or China-specific audience. Under the natural-language policy rule, forcing a specific language without opt-in can be a locale-policy concern.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

This markdown file describes the query as supporting broad natural-language searches for "hotels, flights, etc." and then provides open-ended pattern categories including an "Other" bucket. That scope is ambiguous enough that users or routing systems may not clearly know what queries are in-bounds versus out-of-scope, increasing the risk of unintended invocation or misuse.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

This markdown file includes invocation examples using Chinese city names and output fields/values in Chinese, but it does not tell users that the skill is China/Chinese-locale specific or offer an alternative locale. Under the policy, forcing a specific language or locale without opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The parameter documentation requires hotel type and bed type values in Chinese (酒店, 民宿, 客栈, 大床房, 双床房, 多床房) and the examples also use Chinese destination names. This creates a language/locale constraint in the natural-language interface, but the file does not state that the skill is region-specific or offer an alternative language option.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The documentation requires bed-type values in Chinese and all usage examples use Chinese destination and hotel names, which imposes a specific language/locale convention. There is no note that the skill is China-specific or that users may provide values in other languages, so this appears to be an undocumented locale constraint.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The file includes example input and returned content in Chinese, such as the keyword example and hotel/package names, but does not indicate that the skill is China-specific or that users may choose another language or locale. This can create a language/locale policy issue because the documentation implicitly assumes Chinese-language usage without 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 all user-facing guidance and examples exclusively in Chinese, with no indication that the skill should adapt to the user's preferred language or that the Chinese-only constraint is intentional for a region-specific use case. Under the policy, forcing a specific language without opt-in can be a natural-language policy violation.

Content

No source excerpt is available for this finding.