酒店智能搜索

• Outcome: Pass
• Latest audit: Jun 8, 2026 at 12:29 PM UTC
• Version:
  1.0.0

Overview

This hotel-search skill is purpose-aligned and disclosed, but users should understand that their travel queries are sent through the publisher's cloud proxy to travel and map services.

Install only if you are comfortable sending hotel, destination, and nearby-food queries through the publisher's Tencent Cloud proxy to Fliggy and Gaode. Avoid entering sensitive personal details in searches, and treat prices and booking links as live third-party results that should be verified before purchase.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The skill advertises a narrow 5-tool hotel workflow, but `_build_tips` injects unrelated recommendations such as attractions, train tickets, flights, and city transport into outputs. This is a capability-misrepresentation issue that can mislead users or downstream agents into believing the skill supports functions outside its declared scope, increasing the chance of unintended tool chaining or trust abuse.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The code sends user-provided hotel queries and geolocation-related address data to external proxy endpoints (`FLIGGY_PROXY` and `GAODE_PROXY`) without any visible consent, notice, or data-handling disclosure. Because travel queries can reveal sensitive location, itinerary, or accommodation preferences, undisclosed third-party transmission creates privacy and data-governance risk.

Missing User Warnings

Medium

• Confidence: 99% confidence
• Finding: A hardcoded proxy token is embedded directly in the script and then sent on every outbound request. Embedded secrets are easily exposed through source distribution, logs, or reuse in other environments, allowing unauthorized use of the proxy service and possible abuse of associated quotas or backend access.

VirusTotal

65/65 vendors flagged this skill as clean.