Back to skill

Security audit

asia-arrival-card

Security checks for vulnerabilities and agentic risk

Overview

Review recommended: the skill is a travel-entry form helper, but it handles passport and travel documents and includes automation guidance for official government submissions with insufficient front-loaded privacy and control boundaries.

Install only if you are comfortable with an assistant helping process sensitive travel documents. Keep official submissions manual, verify every field on the government website yourself, avoid sharing email codes, passwords, payment details, or full passport scans unless necessary, and do not rely on the automation script guidance for live government forms without separate safeguards.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (28)

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill claims to provide guidance, but its embedded automation logic goes beyond advisory content and outlines end-to-end submission behavior against government immigration portals. In context, this would cause the agent to solicit passport, identity, travel, and contact data, then potentially act on the user's behalf on third-party sites, creating risks of unauthorized transactions, mishandling of sensitive data, and unsafe external side effects.

Content

No source excerpt is available for this finding.

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 492)May include surrounding context.

md
- **[country-guides.md](reference/country-guides.md)** - 各国详细填写指南

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 496)May include surrounding context.

md
- **[automation-framework.md](reference/automation-framework.md)** - 自动化脚本设计

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The proposed features include auto-login, automatic form filling, one-click submission, and downloading confirmation files, which move the skill into account interaction and transaction automation. That scope expansion is dangerous because it can enable credential handling, unintended submissions, and high-impact actions on government or travel websites beyond the user's informed expectations.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger phrases are broad enough to invoke the skill for generic travel or visa-related queries, increasing the chance the agent requests sensitive identity and itinerary data when the user may only want general information. In this skill, over-broad invocation is more dangerous because the content then steers toward collecting passport and travel details and even suggests automation of official workflows.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill explicitly asks for highly sensitive personal and travel information—passport number, full name, DOB, nationality, email, phone, itinerary, addresses, and visa number—without clear privacy minimization, masking, retention, or user-safety guidance. In an agent setting, this materially increases the risk of oversharing, accidental storage, prompt leakage, or downstream misuse of identity and travel data.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

Labeling the code as 'pseudocode' does not meaningfully reduce risk when it is structured as near-executable automation for navigating official sites, filling fields, submitting forms, and waiting for results. This can facilitate operational misuse by downstream agents or users who treat it as implementation-ready, especially in a domain involving regulated government submissions and sensitive personal data.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

Handling email verification and saving confirmation artifacts expands the skill from informational assistance into account/process control over sensitive immigration workflows. That creates unnecessary exposure to inbox-linked verification codes, confirmation PDFs, QR codes, and travel credentials, all of which are sensitive and could be misused or retained insecurely.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill description and all interaction guidance are written as if the assistant will operate in Chinese only, but there is no indication that users may choose another language. This can violate language/locale policy because it imposes a specific language without opt-in or a documented region-specific justification.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest and skill name position this as a Southeast Asia arrival-card assistant, with the description centered on Malaysia, Thailand, Singapore, Indonesia, Vietnam, the Philippines, Japan, and Korea. However, the instructions explicitly say that for countries outside the listed set, the skill should use web_search to provide arrival-card guidance for any country, which materially broadens the operational scope beyond the declared regional purpose.

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 automatic extraction of highly sensitive passport, travel, and lodging data from uploaded documents, but does not provide explicit privacy, retention, minimization, or consent guidance. In this context, users may upload passports and tickets containing identity data, document numbers, dates of birth, and travel itineraries, creating material privacy and identity-theft risk if the data is stored, echoed, or processed beyond user expectations.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The document provides an automation framework for submitting highly sensitive personal and travel data to government websites, but it does not prominently warn users about the privacy and security risks of transmitting passport details, birth dates, travel itineraries, and email addresses through automated tooling. In this skill context, the risk is elevated because the skill explicitly encourages automation script suggestions for immigration and visa workflows, which may normalize unsafe handling of regulated personal data and lead users to store, log, or transmit that data insecurely.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

This file presents all user-facing instructions solely in Chinese, including critical travel and compliance guidance. Under the stated policy, forcing a specific language without user opt-in can be a natural-language policy violation unless the locale restriction is clearly documented and justified.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The guide gives contradictory instructions about photo upload: earlier it says users must upload a personal photo, while the FAQ later says no passport photo or documents are required. In a travel-entry workflow, inconsistent instructions can cause users to submit wrong or incomplete information, waste time near departure, or disclose extra personal data unnecessarily.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The primary workflow encourages users to upload passports and other highly sensitive travel documents before prominently presenting privacy and risk warnings. This is dangerous because users may disclose identity documents without understanding cloud processing, retention, third-party access, or safer manual-entry alternatives.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The document expands the skill from guidance into document ingestion and cloud-based recognition of passports, tickets, and hotel bookings, which materially increases the data sensitivity and operational scope. For an arrival-card helper, this creates unnecessary exposure of identity and travel data and can lead users to share documents they would not expect to upload for a simple advisory workflow.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The document states that all information 'must' be filled in using English uppercase letters and explicitly marks using Chinese as incorrect. This is a language policy constraint expressed in natural language, but the file does not explain that the requirement comes from the external government form or otherwise justify the restriction, nor does it offer user choice.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
99% confidence
Finding

This section says that unsupported-country handling should explain that the skill currently supports '6' Southeast Asian countries, while the rest of the file consistently lists 8 pre-prepared countries and also includes Japan and Korea. That contradiction can mislead users and downstream agents about the intended supported scope.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

SQP-3 applies to all file types and covers language or locale policy violations. This markdown file presents all instructions only in Chinese, with no indication that users may choose another language or that the content is intentionally limited to a Chinese-speaking audience.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

This markdown file contains user-facing guidance exclusively in Chinese, and there is no indication that other language options are available or that the language restriction is intentional and justified. Under the policy rule for language or locale constraints, forcing a specific language without user 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
88% confidence
Finding

The instruction states that all information should be entered using uppercase English letters, which imposes a language/locale constraint in natural-language guidance. The document does not present this as optional or explain whether this is an official system requirement, so it risks violating the policy against forcing a specific language without opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

SQP-3 applies to all file types and covers language or locale policy violations in natural language. Line L440 instructs users to select '简体中文' for the page, which imposes a specific language preference rather than neutrally offering it as an option for users who may prefer another available language.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The file’s user-facing FAQ content is entirely in Chinese and does not indicate that language selection is optional or that the skill is intended only for Chinese-speaking users. Under the policy, forcing a specific language without user opt-in can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

Line L039 instructs users to select a specific language/localized interface and labels it as recommended, which can be read as forcing a language preference rather than offering a neutral choice. Under the policy, language or locale constraints should be optional or clearly user-selected.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
75% confidence
Finding

The manifest describes a helper for arrival-card/visa application filling guidance, field explanations, process steps, exception handling, and automation suggestions. This file goes beyond form-filling assistance by providing standalone customs declaration thresholds, prohibited-items rules, and health declaration/travel-history compliance guidance that functions as broader border-entry advisory content rather than just explaining how to fill the form.

Content

No source excerpt is available for this finding.