ClawHub
Back to skill

Security audit

Museum Explorer

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed museum-trip journaling assistant that stores visit notes locally and uses public museum sources, with some web-scraping guidance users should keep bounded.

Installers should expect the skill to create local journal folders, update local exhibition data, and browse public museum sources for verification. Avoid using it to probe undocumented or restricted endpoints, and keep any photo inputs local as the skill describes.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Natural-Language Policy Violations

Medium
Confidence
91% confidence
Finding
The skill includes an explicit instruction to the LLM about how to update files and respond, which can steer behavior without checking the user's current language or formatting preference. While the content is a travel journal rather than a high-risk domain, forcing a specific interaction style or locale can reduce user control and cause the agent to ignore user-specified preferences.

Ssd 2

Medium
Confidence
87% confidence
Finding
The document explicitly encourages discovery of WeChat article links via generic search engines and describes a major museum endpoint as an 'interface already cracked,' which normalizes bypass-oriented collection of content and use of unofficially exposed endpoints. Even if intended for scraping public museum data, this guidance can push agents toward accessing non-documented routes and evading normal access patterns, increasing legal, policy, and abuse risk.

Ssd 2

Medium
Confidence
90% confidence
Finding
The instruction to inspect page JavaScript for $.get/fetch calls and hunt for 'api/search/json' behind SPA shells is active guidance for semantic probing of hidden backend interfaces. In a scraping skill, that materially increases the chance the agent will use unintended, undocumented, or weakly protected endpoints, which can violate access expectations and be repurposed for broader reconnaissance.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.