← Back to Skills Marketplace
188
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install nvidia-smi
Description
Provides detailed guidance on using the nvidia-smi command for real-time NVIDIA GPU monitoring, management, troubleshooting, and automation.
Usage Guidance
This is a coherent, instruction-only tutorial for nvidia-smi and NVML. Before installing or enabling it: 1) Confirm your system actually has NVIDIA drivers and nvidia-smi available; the skill assumes these tools exist. 2) Be aware the instructions include process termination (kill <PID>) — only allow the agent to run such commands if you trust it or review actions first. 3) The Python example requires the pynvml library; install it manually (pip install nvidia-ml-py3 or similar) if you plan to run the sample. 4) Because the skill can instruct shell actions, consider limiting autonomous execution or requiring user confirmation for destructive commands. No credentials are requested by the skill, which reduces risk.
Capability Analysis
Type: OpenClaw Skill
Name: nvidia-smi
Version: 1.0.0
The skill bundle provides legitimate documentation and a Python script example for using the 'nvidia-smi' utility to monitor and manage NVIDIA GPUs. The content in SKILL.md is purely educational and functional, focusing on standard system administration tasks like monitoring VRAM, temperature, and process management without any indicators of malicious intent or data exfiltration.
Capability Assessment
Purpose & Capability
The SKILL.md content matches the stated purpose: detailed guidance for nvidia-smi, NVML/pynvml usage, monitoring, troubleshooting, and basic process management. It references expected tools (nvidia-smi, watch, kill, pynvml). The registry metadata declared no required binaries — a minor documentation gap but not an incoherence: the referenced commands are appropriate to the stated purpose.
Instruction Scope
Instructions stay within the GPU monitoring/management scope. They advise running nvidia-smi, using watch/dmon, parsing CSV output, and include a Python pynvml monitoring example. They also recommend killing processes (kill <PID>) to free VRAM, which is a legitimate but potentially destructive admin action; the skill does not instruct reading unrelated files or exfiltrating data.
Install Mechanism
This is an instruction-only skill with no install spec or code files, which is the lowest-risk install model. The Python example references the external library 'pynvml' but no install instructions are provided — this is an omission for usability but not an active security risk from the skill bundle itself.
Credentials
The skill requests no environment variables, credentials, or config paths. It performs tasks that do not require secrets. This is proportionate to the stated purpose.
Persistence & Privilege
The skill is not marked always:true and does not request persistent system-level privileges. Note: the skill's runtime instructions include commands that terminate processes (kill) and run long-running monitoring; if the agent is allowed to execute shell commands autonomously, that capability could be used to perform destructive actions. Autonomous invocation itself is the platform default and is not a fault of the skill, but users should be aware.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install nvidia-smi - After installation, invoke the skill by name or use
/nvidia-smi - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release: Comprehensive guide to mastering the `nvidia-smi` command for GPU monitoring, management, and troubleshooting.
- Covers basic and advanced `nvidia-smi` usage for status checks, process management, and detailed queries.
- Explains interpretation of key output fields, including memory usage, GPU utilization, and process information.
- Provides actionable workflows for diagnosing "CUDA out of memory" errors and performance bottlenecks.
- Introduces continuous monitoring, filtering, and daemon mode for real-time updates.
- Includes automation tips, with example Python monitoring script using the `pynvml` library for direct NVML access.
Metadata
Frequently Asked Questions
What is Nvidia smi tutorial?
Provides detailed guidance on using the nvidia-smi command for real-time NVIDIA GPU monitoring, management, troubleshooting, and automation. It is an AI Agent Skill for Claude Code / OpenClaw, with 188 downloads so far.
How do I install Nvidia smi tutorial?
Run "/install nvidia-smi" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Nvidia smi tutorial free?
Yes, Nvidia smi tutorial is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Nvidia smi tutorial support?
Nvidia smi tutorial is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Nvidia smi tutorial?
It is built and maintained by wow (@duanc-chao); the current version is v1.0.0.
More Skills