ClawHub
Back to skill

Security audit

网红水族馆

Security checks for vulnerabilities and agentic risk

Overview

This Markdown-only skill should be reviewed because it claims to guide aquarium visits but mainly gives trail and route-planning instructions.

Install only if you are comfortable reviewing mismatched guidance. The publisher should align the skill so it is clearly either an aquarium guide or a route/trail planner before users rely on it for visit planning, reservations, safety, or logistics.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The manifest description says the skill handles '网红水族馆' recommendations and visit guides, but the documented behavior is clearly for outdoor trail/travel routing. This creates a semantic integrity failure: the orchestrator may route aquarium-related user queries into a skill that produces unrelated route-planning advice, undermining trust and potentially leading to unsafe decisions if users rely on mismatched safety, booking, or travel information.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The activation wording is broad enough that the skill could be invoked for loosely related travel-planning requests without clear boundaries. When combined with the domain mismatch in this file, overbroad triggering increases the chance of incorrect tool selection and irrelevant or misleading responses.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill is labeled as an aquarium recommendation/guide, but the body content describes hiking-style route planning with terrain, elevation, and trail safety fields. This mismatch can cause the agent to invoke the skill in the wrong context and return irrelevant or unsafe travel guidance to users expecting aquarium information.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
68% confidence
Finding

All user-facing content in the skill file is written in Chinese, and there is no indication that users may choose another language or that the locale restriction is intentional and documented. This may conflict with language or locale policy expectations if the skill is exposed to a broader user base.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.