ClawHub
Back to skill

Security audit

Cruise Travel Advice

Security checks for vulnerabilities and agentic risk

Overview

This skill is a sparse remote cruise-planning MCP integration, but its disclosed behavior is related to its stated purpose and shows no evidence of hidden or harmful actions.

Install only if you are comfortable sending cruise-related requests and product identifiers to the CruiseSkillBridge/olavacations remote MCP gateway. Avoid including secrets, payment details, account credentials, or unnecessary personal information.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The skill is described as providing cruise travel advice, but the published MCP tool exposed in the document is `product_info`, which retrieves product details by product number. This mismatch can mislead users and downstream agents about the skill’s actual capabilities, causing unintended data access or invocation of a tool outside the user’s expected task scope.

Intent-Code Divergence

Medium
Confidence
94% confidence
Finding
The front-matter description advertises travel-advice functionality, while the only concrete tool documented later performs product lookup. This kind of capability misrepresentation is dangerous in an agent ecosystem because routing, trust decisions, and user consent may be based on the declared description rather than the actual callable behavior.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The documentation instructs sending POST requests to an external endpoint and states that requests will be forwarded through a gateway and counted in console statistics, but it does not warn users that their inputs are transmitted to third parties and may be logged or tracked. In an agent setting, this creates a privacy and data-governance risk because sensitive prompts or identifiers could be exfiltrated outside the local trust boundary without clear disclosure.

Static analysis

No suspicious patterns detected.