Hardware Llm Optimizer

Auto-detect PC hardware (CPU/GPU/RAM/VRAM) -> Determine max LLM parameters -> Recommend models (3B/7B/8B/13B/34B/70B) + quantization + deployment tools + bot...

Security Scan

The skill's code and runtime instructions match its stated purpose (local hardware detection and LLM recommendations); it only reads local system info and does not request credentials, external installs, or network endpoints.

Purpose & Capability

Name/description (detect hardware, recommend models/quantization/deployment) align with included files: SKILL.md instructs running detect.py and detect.py inspects CPU, RAM, GPU and gives model/quantization recommendations. No unrelated credentials, binaries, or services are requested.

Instruction Scope

SKILL.md directs the agent to run the bundled detect.py. The script only reads local system info (psutil), checks /proc/version for WSL, and calls nvidia-smi for GPU details — all within the stated purpose. It does not access external endpoints, env vars, or other user files beyond /proc/version.

Install Mechanism

No install spec; this is instruction-only with a bundled Python script. The only required third-party package is psutil (documented in SKILL.md). The script does not download or execute code from remote URLs.

Credentials

No environment variables, credentials, or config paths are requested. The script uses local APIs and optional nvidia-smi; this is proportionate to hardware-detection functionality.

Persistence & Privilege

Skill is not always-enabled, does not attempt to modify other skills or agent-wide settings, and does not persist credentials. It merely prints local detection results.

Assessment

This skill appears to do what it claims: run locally, detect hardware, and print recommendations. Before installing/running: (1) review the detect.py yourself (it's included) and confirm you're comfortable executing it, (2) be aware it will reveal local hardware details (CPU/GPU/RAM) to whatever component runs the skill — if your agent forwards outputs to external services, those details could be transmitted, and (3) install psutil (pip install psutil) and ensure nvidia-smi is available if you want GPU detection. If you want extra caution, run the script in a local sandbox or VM first.

Detects PC hardware configuration and recommends which large language models can run.

Features

1. Auto-detect: CPU, RAM, GPU (NVIDIA/AMD), VRAM
2. Calculate: Maximum runnable model size
3. Quantization: FP16 / 8bit / 4bit / 2bit recommendation
4. Model Suggestion: Llama 2/3, Qwen, Mistral, Phi, Gemma, Yi, etc.
5. Bottleneck Analysis: System constraint diagnosis
6. Deployment Tools: Ollama, Llama.cpp, vLLM, Chatbox
7. Optimization Tips: Low VRAM solutions
8. Minimum Config Table: 3B/7B/13B/34B/70B requirements

Usage

When user asks about running LLMs on their computer:

检测电脑配置
大模型推荐
能跑什么模型
硬件检测
LLM优化

Quick Run

python3 skills/hardware-llm-optimizer/detect.py

Requirements

• Python 3.8+
• psutil: pip install psutil
• nvidia-smi (optional, for NVIDIA GPU detection)

Minimum Config Reference

Model	Min VRAM	Rec VRAM	Quantization
3B	2GB	4GB	Q4
7B	6GB	8GB	Q4/Q8
13B	10GB	16GB	Q4/Q8
34B	20GB	32GB	Q4
70B	40GB	80GB	Q4

Chinese Interface

This skill outputs in Chinese for user convenience.