ClawHub
Back to skill

Security audit

Mental Models Master | 顶级思维模型工具箱

Security checks across malware telemetry and agentic risk

Overview

This is a text-only decision framework that may over-apply investment-style analysis, but it does not request access, credentials, persistence, or authority to act.

Install this only if you want the assistant to use structured mental-model analysis for important decisions. For investing, verify facts independently, consider qualified financial advice, and keep buy/sell decisions under your own control.

SkillSpector

By NVIDIA
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 (3)

Description-Behavior Mismatch

Medium
Confidence
89% confidence
Finding
The skill metadata presents a broad 'thinking-model toolbox,' but the body operationalizes it as an investment-decision assistant with explicit buy/sell, valuation, and portfolio guidance. This mismatch can cause the host agent or user to invoke the skill under a safer-sounding label and receive domain-specific financial guidance without appropriate scoping, warnings, or policy controls.

Description-Behavior Mismatch

Medium
Confidence
87% confidence
Finding
The skill claims applicability to 'any important problem/decision,' yet its triggers and content are predominantly investment-oriented. That overclaim increases the chance of the skill being activated in unrelated contexts and steering conversations toward financial framing or advice that is inapplicable or risky.

Vague Triggers

High
Confidence
94% confidence
Finding
The trigger conditions are extremely broad, including phrases like 'how to analyze,' 'what is worth it,' and 'what should I do,' which are common in many benign conversations. In an agent setting, this can cause frequent unintended activation, override user intent, and inject unnecessary or inappropriate decision-making behavior into unrelated tasks.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.