ClawHub
Back to skill

Security audit

Hardware LLM Optimizer v2 (llmfit)

Security checks across malware telemetry and agentic risk

Overview

This appears to be a legitimate hardware-based LLM recommendation skill, but it includes an unverified remote installer that can execute code on the user's machine.

Review the installer path carefully before use. Prefer Homebrew, uv/pip, Docker, or a manually downloaded release with checksum or signature verification instead of running the curl-piped shell command. Expect the tool to inspect and display hardware details such as CPU, RAM, GPU, VRAM, and driver information.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (5)

Lp3

Medium
Category
MCP Least Privilege
Confidence
77% confidence
Finding
The skill exposes shell-executable commands and appears capable of reading local system state, yet it declares no explicit permissions. That creates a trust and review gap: users or orchestrators may invoke hardware detection and shell commands without realizing the skill needs elevated capabilities, increasing the chance of unintended command execution or host data exposure.

Tp4

High
Category
MCP Tool Poisoning
Confidence
89% confidence
Finding
The skill advertises broad hardware detection, simulation, TUI support, speed estimation, and a large model library, but the documented behavior does not substantiate those claims. Security-wise, this is dangerous because overstated capabilities can mislead users and reviewers into trusting outputs or approving execution paths that are not actually implemented or validated.

Vague Triggers

Medium
Confidence
73% confidence
Finding
The trigger phrases are very broad, such as generic questions about recommended LLMs or hardware detection, which may cause the skill to activate during ordinary conversation without clear user intent. In a skill that can lead to shell usage and installation guidance, over-broad activation increases the risk of unsolicited command suggestions or execution paths being introduced into benign chats.

External Script Fetching

High
Category
Supply Chain
Content
## 安装llmfit（如需）

```bash
curl -fsSL https://llmfit.axjns.dev/install.sh | sh
```

## 优势对比
Confidence
99% confidence
Finding
curl -fsSL https://llmfit.axjns.dev/install.sh | sh

Chaining Abuse

High
Category
Tool Misuse
Content
## 安装llmfit（如需）

```bash
curl -fsSL https://llmfit.axjns.dev/install.sh | sh
```

## 优势对比
Confidence
98% confidence
Finding
| sh

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.