Tokyo

• Outcome: Pass
• Latest audit: Sep 10, 2026 · 8:45 PM UTC
• Version: 1.0.0

Overview

This is a markdown-only Tokyo guide with travel, relocation, culture, and local-practical advice, and it does not request code execution, credentials, persistence, or privileged access.

Reasonable to install as a Tokyo reference skill. Users should treat prices, visa notes, business data, and current travel details as guidance that may need verification, and some cultural or language-accessibility wording would benefit from more nuance.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (14)

Anti-Refusal Statement

High

• Category: Anti-Refusal

Content

### Cultural Bridge

**Best approach**: Respectful curiosity without judgment. Japan does things differently, not wrong.

• Confidence: 80% confidence
• Finding: Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Vague Triggers

Medium

• Confidence: 96% confidence
• Finding: The activation rule is extremely broad: any user asking about Tokyo 'for any purpose' will invoke this skill. Overly broad triggering can cause the agent to select this skill in loosely related contexts, increasing the chance of irrelevant guidance, policy bypass through unintended tool routing, or suppression of more appropriate specialized skills. In this case the content is informational and not inherently dangerous, so the main risk is misrouting rather than direct exploitation.

Natural-Language Policy Violations

Low

• Confidence: 85% confidence
• Finding: Lines L309-L310 recommend apps labeled explicitly as 'Japanese' and 'Japanese, accurate' without clarifying that these resources may be unsuitable for users who do not read Japanese. This can create a locale/language policy issue because the content steers users toward a specific language experience without offering a choice or opt-in.

Natural-Language Policy Violations

Low

• Confidence: 84% confidence
• Finding: The pricing table labels haircut costs as 'Haircut (men)' and 'Haircut (women)', which imposes a gendered categorization in natural-language content. This can conflict with organizational language policy expectations where user-facing content should avoid forcing identity-based categories when neutral alternatives are possible.

Natural-Language Policy Violations

Low

• Confidence: 79% confidence
• Finding: Phrases like 'The foundation of Japanese society' and 'Japanese people may say' present broad cultural generalizations as universal rules. Because this file is guidance text, the lack of qualifiers such as 'often,' 'in some contexts,' or 'may vary by region/person' can push users toward an overly rigid locale-specific interpretation.

Natural-Language Policy Violations

Low

• Confidence: 72% confidence
• Finding: This wording is an unconditional cultural directive in natural language. While tipping is generally uncommon in Japan, expressing it as an absolute without noting exceptions or context can violate the policy against forcing rigid locale behavior without justification or user choice.

Natural-Language Policy Violations

Low

• Confidence: 69% confidence
• Finding: The section frames users through a fixed outsider label and includes a colloquial term that can be perceived as insensitive. In policy terms, this is a natural-language locale framing issue because it assigns a cultural identity and interaction model without qualification.

Natural-Language Policy Violations

Low

• Confidence: 89% confidence
• Finding: The text makes a broad assertion that outside Minato ward users should assume no English support, which can mislead users and encode an overgeneralized stereotype rather than present situational guidance. In a location-advice skill, this is not a code-execution or direct security flaw, but it is still a quality/safety issue because users may make decisions based on inaccurate or exclusionary assumptions.

Natural-Language Policy Violations

Low

• Confidence: 79% confidence
• Finding: The phrase "English limited despite tourists" is a broad language-accessibility claim that can influence user decisions about where to live or visit without context, nuance, or user preference. In a neighborhood guidance skill, this can steer non-Japanese-speaking users away from an area based on a generalized accessibility judgment rather than concrete, opt-in criteria such as multilingual services, signage, or transit support.

Natural-Language Policy Violations

Low

• Confidence: 83% confidence
• Finding: Line L005 contains Japanese characters in parentheses, which introduces a specific language without any user opt-in or explanation. Under the stated policy, forcing or assuming a language/locale can be a natural-language policy issue when no choice is provided.

Natural-Language Policy Violations

Low

• Confidence: 81% confidence
• Finding: This markdown content presents all guidance, headings, and descriptors in English for a location-specific housing guide, with no indication that English is optional or that other language preferences are supported. Under the policy, forcing a specific language without user opt-in can be a natural-language policy issue.

Natural-Language Policy Violations

Low

• Confidence: 76% confidence
• Finding: Line L073 states 'Limited English support,' which introduces a language-related constraint in the skill content. While descriptive, it presents a locale/language limitation without clarifying alternatives, user choice, or why the constraint is necessary.

Natural-Language Policy Violations

Low

• Confidence: 81% confidence
• Finding: The guide specifically highlights 'English support limited' and an 'English medical help' line, which centers a single language rather than offering broader language/locale options. Under the policy rule, forcing or presuming a specific language without user opt-in can be a natural-language policy concern.

Natural-Language Policy Violations

Low

• Confidence: 78% confidence
• Finding: The line states 'Japanese required' as a blanket characteristic, which is a language/locale constraint presented without user choice or an explicit policy justification. While the document is informational, this wording can be read as enforcing a specific language expectation rather than neutrally describing context.

Static analysis

No suspicious patterns detected.