网红水族馆

• Outcome: Review
• Latest audit: May 29, 2026 at 9:27 PM UTC
• Version:
  0.1.0

Overview

This instruction-only skill has no code-level access risk, but its aquarium label conflicts with hiking-style route and safety guidance that could mislead users.

Review before installing. This skill does not appear to run code or access data, but its content should be corrected so aquarium guidance focuses on exhibits, tickets, hours, transit, accessibility, crowds, and venue rules rather than hiking-route criteria.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Intent-Code Divergence

High

• Confidence: 98% confidence
• Finding: The skill’s body documents hiking and outdoor route-planning features such as mileage, elevation, road conditions, emergency retreat, and trail logistics, which materially conflict with the declared aquarium purpose. This kind of semantic mismatch can cause the wrong skill to be invoked for user requests, leading to unsafe or misleading guidance and breaking routing and trust assumptions in the agent ecosystem.

Description-Behavior Mismatch

High

• Confidence: 97% confidence
• Finding: The documented behavior clearly describes outdoor hiking/trail guidance rather than aquarium recommendation and visit assistance. A mis-scoped skill can be selected in inappropriate contexts and provide irrelevant or unsafe advice, especially where users rely on the assistant to match domain-specific planning and safety considerations correctly.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The activation description is broad enough to match generic travel-planning or itinerary requests, not just aquarium-related queries. Overbroad triggering increases the chance of accidental invocation, causing cross-domain interference and exposing users to the wrong workflow or recommendations.

VirusTotal

64/64 vendors flagged this skill as clean.