ClawHub
Back to skill

Security audit

pricewin-hotel-search

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only hotel search skill that uses a disclosed PriceWin MCP backend and does not install code, persist, or transact.

Install this only if you are comfortable sending city, dates, guest count, language code, and polling session IDs to PriceWin's hosted MCP backend. Be aware that results may default to Vietnamese because the skill hard-codes that language parameter.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

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 (2)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill instructs the agent to always send language="vi" when calling the hotel search tool, regardless of the user's actual language or locale. This can cause unintended disclosure of inferred locale preferences, produce mismatched search results or localized pricing/content, and override user intent without consent or justification.

Content

No source excerpt is available for this finding.

Behavior Manipulation

Medium
Category
Prompt Injection
Confidence
77% confidence
Finding

The instruction to continue polling silently and avoid telling the user that work is still in progress is a form of behavior manipulation that reduces transparency about ongoing backend activity. In this skill's context, that matters because queries are being sent to a hosted, closed-source intermediary, so suppressing status disclosure can hide extended external processing and network use from the user.

Content

Scanner excerpt · SKILL.md (reported line 37)May include surrounding context.

md
4. Present results as soon as status == "partial" with hotels
5. Continue polling silently — refine if more arrive

**Never tell the user "loading/please wait" after 1-2 polls — that's premature.**

## 4th source: OpenTravel