Amsterdam

• Outcome: Pass
• Latest audit: Sep 10, 2026 · 8:50 PM UTC
• Version: 1.0.1

Overview

This is a static Amsterdam guide with no executable behavior, credentials use, persistence, or hidden authority, though users should verify time-sensitive health, visa, legal, and cost details.

Installers should treat this as a general Amsterdam reference, not an authoritative source for medical, immigration, tax, insurance, or legal decisions. Verify current official details before acting, especially because the skill labels some data as February 2026 and the review date is September 10, 2026.

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

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• YARA Signatures: Malware Match, Webshell Match, Cryptominer Match
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (15)

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The phrase "User asks about Amsterdam for any purpose" is extremely expansive and does not define clear constraints or exclusion conditions. Because this is a markdown skill description, the trigger guidance should be more specific to avoid unintended invocation on generic conversation about Amsterdam.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The file provides detailed healthcare guidance, costs, care pathways, emergency routing, and insurance advice without any disclaimer that information may change, be incomplete, or not substitute for professional medical or legal/insurance advice. In a healthcare context, outdated or overgeneralized guidance can cause users to delay urgent care, misunderstand coverage obligations, or make unsafe decisions based on incomplete information.

YARA rule 'network_reconnaissance': Network reconnaissance and scanning patterns [hacktools]

Medium

• Category: YARA Match

Content

d market lovers
- Young professionals
- Those who don't mind crowds

**Oud-Zuid:**
- Families with budget
- Culture seekers
- Those wanting prestige
- Professionals seeking quality

**Rivierenbuurt:**
- Families wanting value
- Architecture lovers
- Those seeking quiet
- Professionals with young families

---

*South Amsterdam represents the city's most desirable residential areas. Competition is fierce, so prepare thoroughly and act fast.*

• Confidence: 65% confidence
• Finding: YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Natural-Language Policy Violations

Medium

• Confidence: 87% confidence
• Finding: This markdown file contains natural-language guidance that privileges English as sufficient for work in Amsterdam. Under the language/locale policy rule, statements that force or assume a specific language without user opt-in can be policy violations, especially when no alternative language options or caveats are offered.

Natural-Language Policy Violations

Medium

• Confidence: 92% confidence
• Finding: This line makes a categorical language claim that centers one language choice and may exclude users who need locale-sensitive guidance. The file does not present this as optional or context-dependent, so it fits the language/locale policy concern.

Natural-Language Policy Violations

Low

• Confidence: 77% confidence
• Finding: These lines instruct the user to adopt a specific communication style such as being direct and skipping excessive politeness, framed as generally applicable cultural guidance. Because the advice is presented as a blanket norm rather than optional or context-dependent, it risks imposing a single interaction style without opt-in or nuance.

Natural-Language Policy Violations

Low

• Confidence: 77% confidence
• Finding: Lines L043-L046 instruct the reader to 'Learn Dutch,' 'Be direct,' and 'Accept their style,' which frames assimilation into a specific language and social style as expected rather than optional. Because SQP-3 covers language or locale policy violations, this can be seen as steering users toward a specific locale behavior without explicit choice or qualification.

Intent-Code Divergence

Low

• Confidence: 92% confidence
• Finding: At L142 the heading says 'Utilities (DEWA)', but the surrounding content lists Dutch/Amsterdam utility providers like Vattenfall, Eneco, and Waternet. DEWA is associated with Dubai utilities, so the heading actively conflicts with what the section actually covers and could mislead users about the jurisdiction.

Natural-Language Policy Violations

Low

• Confidence: 94% confidence
• Finding: The phrase 'Language barriers in some shops' frames limited English accessibility as a neighborhood drawback without user-specific context or justification. In a city-guide skill, this can reinforce bias against multilingual communities and steer users away from areas based on perceived linguistic or cultural difference rather than objective housing or transit factors.

Natural-Language Policy Violations

Low

• Confidence: 78% confidence
• Finding: The section states "Less than US/UK" and describes a "generally privacy-respecting culture," which presents a broad comparative policy claim tied to locale/culture without qualification or user choice. This can be read as imposing an unqualified regional framing rather than neutrally presenting facts.

Natural-Language Policy Violations

Low

• Confidence: 78% confidence
• Finding: Line L221 says 'English works for most roles' as a general hiring reality. This can be read as privileging a specific language without clarifying that language needs may vary by role, company, or user preference, which may conflict with a language/locale choice policy.

Intent-Code Divergence

Low

• Confidence: 97% confidence
• Finding: At L402, the key takeaway says 'Family comes free,' which contradicts the documented costs in L313-L318 showing application fees for spouse/partner and child. This is an intent/documentation contradiction within the file, not merely an omission.

Missing User Warnings

Low

• Confidence: 93% confidence
• Finding: The file advises users to rent a bike without any accompanying safety context, despite Amsterdam traffic posing well-known risks to inexperienced cyclists, especially around trams, dense bike lanes, and pedestrians. In a visitor-oriented travel skill, this omission can contribute to unsafe behavior by tourists who may treat cycling as casual sightseeing rather than a transport activity requiring local rules and caution.

Missing User Warnings

Low

• Confidence: 92% confidence
• Finding: The itinerary recommends renting a bike and exploring the city by bicycle without warning about Amsterdam's heavy bike traffic, tram tracks, right-of-way expectations, and local cycling rules. For tourists, this omission can increase the likelihood of crashes or unsafe road behavior.

Missing User Warnings

Low

• Confidence: 90% confidence
• Finding: The nightlife itinerary explicitly centers bars, clubs, and drinking but provides no safety guidance about intoxication, late-night transport, personal security, or staying with companions. In a travel-planning skill, this omission can contribute to risky user behavior, especially for visitors unfamiliar with local areas and norms.

Static analysis

No suspicious patterns detected.