ClawHub
Back to skill

Security audit

网红游戏厅

Security checks for vulnerabilities and agentic risk

Overview

The skill is a simple Markdown-only recommendation guide, but its title and description say game halls while the actual criteria and examples point to fashion, cosplay, and handmade retail recommendations.

Review this skill carefully before installing because it appears low-risk technically, but its purpose is unclear. It should be corrected to consistently cover either game halls or fashion/handmade retail so users and agents know when to invoke it.

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
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill’s declared purpose is recommending internet-famous game halls, but its filters and returned fields describe fashion/cosplay retail selection criteria such as authenticity, limited editions, try-on rules, and makeover services. This kind of domain mismatch can cause the orchestrator to invoke the skill for unrelated user intents, leading to deceptive outputs, unsafe routing, and erosion of trust in downstream recommendations.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The example requests ask for trendy fashion/handcraft lists rather than game-hall recommendations, directly contradicting the stated skill purpose. Examples strongly influence invocation and usage patterns, so contradictory examples can misroute user requests, produce irrelevant results, and conceal the skill’s real behavior from reviewers and systems.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.