nvidia_gpu_exporter AGENTS.md

nvidia_gpu_exporter AGENTS.md is an instructions file for Codex, OpenCode from utkuozdemir/nvidia_gpu_exporter. It costs 6,168 tokens per session, scanned A, original, MIT.

A repository guide for contributors and coding agents working on a tool that collects NVIDIA GPU metrics. It explains the code layout, collection methods, test data, and testing rules.

In plain words
What is it for?
Use it to understand how the NVIDIA GPU exporter is organised, where its data comes from, and how to test changes.
Why use it?
It keeps important project knowledge in one maintained place and helps prevent agents from relying on outdated lists or assumptions.

Instructions file for CodexOpenCode

About the project

nvidia_gpu_exporter is a monitoring exporter that reads NVIDIA GPU measurements through the nvidia-smi program or the NVML driver library and exposes them to Prometheus. It is used to monitor GPU utilization, memory, power, and temperature on machines, containers, Kubernetes clusters, and other supported setups. The catalogue instructions cover deploying or configuring this exporter.

utkuozdemir/nvidia_gpu_exporter · 1,550 stars · on GitHub

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/utkuozdemir/nvidia_gpu_exporter/agents-md
Clone the repo
git clone --depth 1 https://github.com/utkuozdemir/nvidia_gpu_exporter

Made for: Codex, OpenCode.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for nvidia_gpu_exporter AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/utkuozdemir/nvidia_gpu_exporter/agents-md.svg)](https://agentmods.dev/instructions/utkuozdemir/nvidia_gpu_exporter/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/utkuozdemir/nvidia_gpu_exporter/agents-md"><img src="https://agentmods.dev/badge/instructions/utkuozdemir/nvidia_gpu_exporter/agents-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 6,168 This file is loaded in full into every session.
When invoked 6,168 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.06168 $0.06168
Opus 5 $0.03084 $0.03084
Sonnet 5 $0.01234 $0.01234
Haiku 4.5 $0.00617 $0.00617

Measured yesterday against content hash cc206a0b268f, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

nvidia_gpu_exporter AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 304 lines

How it starts

The opening of the file, as written. The whole thing — 304 lines — stays where its author put it; the contents beside it link to each section on GitHub.

nvidia_gpu_exporter — agent guide

This is the project's knowledge base for humans and agents working in this repository. Maintain it as you go: whenever you learn something durable about how this repo works, add it here. It is not append-only, so fix or delete anything that becomes wrong or outdated.

Keep it timeless. Two rules keep it from rotting: prefer pointers over copies, and prefer invariants over inventories. Never restate a flag list, a metric list, a chart value, a file count, a version number or the current state of some in-flight work. Those live in the code and the docs, they change constantly, and a stale copy here is worse than no copy. What belongs here is the map, and the reasoning that is not written down anywhere else.

CLAUDE.md, GEMINI.md and .github/copilot-instructions.md point at this file, so this is the one place to edit. The Copilot file additionally inlines a few of the rules below, because its consumer injects that file verbatim instead of following the pointer; keep those copies in sync.

Markdown note: task lint:assets runs markdownlint over every tracked markdown file except the generated chart README, this one included, so it has to pass. Keep one sentence per line and the first line as the single top-level heading.

What this repo is

A Prometheus exporter for NVIDIA GPUs, distributed as a single static Go binary. By default it runs nvidia-smi, parses its CSV output and exports the result, which is why it works anywhere that binary works (Linux, Windows, macOS, containers, WSL2) with no C bindings and no NVIDIA userspace of its own.

The distribution surface is wide for a project this size: archives, native packages, container images on more than one registry, a Helm chart, Windows package manifests, a systemd unit, a native Windows service, and the Grafana dashboards. .goreleaser.yml and .github/workflows/release.yml are the authority on what is published and where. Almost all of it is produced by the release pipeline from this repository, with one exception: the dashboards are published to grafana.com by hand.

Read the full file on GitHub · 304 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 304 lines · 6,168 tokens per session scan A cc206a0b268f

Subscribe to this mod's changes

nvidia_gpu_exporter AGENTS.md is an instructions file published in the GitHub repository utkuozdemir/nvidia_gpu_exporter (1,550 stars, last pushed yesterday), licensed MIT. It adds 6,168 tokens to every session, about $0.0308 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.