К основному содержанию
скилл CX Codex

nvshmem-install — скилл для Claude Code: установка NVSHMEM

Скилл nvshmem-install помогает подобрать и проверить установку NVSHMEM или NVSHMEM4Py для пакета, контейнера либо сборки из исходников. Пригодится перед запуском на рабочей станции или кластере.

Apache-2.0 ★ 4 106 обновлён 2026-10-23

Установить в Claude Code · Через Skills CLI (Vercel) Все способы и платформы ↓
npx skills add NVIDIA/skills --skill nvshmem-install -a claude-code
Открыть на GitHub

Что делает nvshmem-install

Скилл собирает требования к установке: нужный интерфейс, версию, тип системы, область установки и способ развёртывания. Затем он сверяет их с документацией для выбранной версии и готовит команды с проверками. Можно запросить NVSHMEM для C/C++, Python через NVSHMEM4Py или оба интерфейса.

Для точных требований и команд нужен скилл $nvshmem-docs в реестре или файл ../nvshmem-docs/SKILL.md. Проверка системы выполняется только после выбора пользователем способа: запустить пробу самому или попросить агента. Проба доступна лишь через непривилегированную оболочку и не подтверждает состояние всего кластера. Это важно для мног узловой установки.

По умолчанию скилл выдаёт команды, но не выполняет их. Перед установкой он должен показать точный блок команд и получить отдельное явное разрешение; права администратора или sudo для проверки системы не используются. Сбой установки останавливает процесс на первой ошибке.

Какие задачи решает

  • подобрать способ установки NVSHMEM под систему и область установки
  • сверить версию NVSHMEM с требованиями документации
  • подготовить команды установки NVSHMEM4Py
  • проверить установку и подготовить двухпроцессную проверку с выбранным лаунчером

Примеры запросов агенту

  • Подбери способ установки NVSHMEM для Python на одном компьютере; сначала задай вопросы о целевой системе и лаунчере.
  • Подготовь установку NVSHMEM версии latest в мой префикс и покажи команды, не выполняя их.
  • Проверь, какие сведения нужны для установки NVSHMEM из исходников, и дождись моего разрешения перед запуском команд.

Как установить nvshmem-install

1Через Skills CLI (Vercel)

npx skills add NVIDIA/skills --skill nvshmem-install -a claude-code

2Вручную: положить файлы

~/.claude/skills/nvshmem-install/SKILL.md

Папка скилла целиком; для одного проекта — .claude/skills/nvshmem-install/

По шагам, с командами для каждой платформы: Как установить скилл в Claude Code: пять способов

Вопросы о nvshmem-install

Как установить nvshmem-install в Claude Code?

Выполните npx skills add NVIDIA/skills --skill nvshmem-install -a claude-code. Файл скилла находится по пути ~/.claude/skills/nvshmem-install/SKILL.md.

Что делает nvshmem-install?

Он помогает спланировать установку NVSHMEM и NVSHMEM4Py, подобрать метод по требованиям системы и подготовить проверки. Установочные команды не запускаются без отдельного разрешения.

Что нужно для точных команд установки?

Нужен $nvshmem-docs в реестре скиллов или файл ../nvshmem-docs/SKILL.md: по ним скилл сверяет требования конкретной версии. Проверка целевой системы возможна только через непривилегированную оболочку.

Исходный текст

SKILL.md · Apache-2.0
---
name: nvshmem-install
description: Plan and validate NVSHMEM and NVSHMEM4Py installations. Use for package, container, or source deployments.
license: Apache-2.0
metadata:
  version: "1.0.0"
  author: NVIDIA NVSHMEM Team <[email protected]>
  tags:
    - nvshmem
    - install
    - deployment
---

# NVSHMEM Installation

## Purpose

Guide safe, version-aware NVSHMEM and NVSHMEM4Py installations on workstations, clusters, multi-node systems, and containers. Select an installation method from documented requirements and observed system facts, then provide exact commands and validation steps. Execute installation actions only after separate explicit approval.

## Prerequisites

- `$nvshmem-docs` must be available in the skill registry or readable at `../nvshmem-docs/SKILL.md` to resolve official, version-specific requirements. If neither route is available, exact requirements and installation commands are blocked; return the mapped official links and identify the unresolved claims. No API keys are required.
- Use registered `$nvshmem-troubleshoot-and-report-bugs` for post-install launch, initialization, runtime, or validation failures. If it is unavailable, preserve the diagnostic evidence and provide official support guidance instead of assuming that an unreadable local fallback exists.
- Confirm network access for remote artifacts and write authority for the chosen installation prefix before execution.
- Do not inspect the target automatically. First complete the installation intake, then ask whether the user wants to run the read-only probe themselves or wants the agent to run it. Run it only after the user explicitly selects the latter.
- Resolve NVSHMEM, CUDA, driver, compiler, launcher, bootstrap, and transport dependencies from the exact-version documentation and target inspection instead of treating them as fixed prerequisites.
- Inspect the target only through an already available unprivileged shell. If direct inspection is unavailable, provide the probe for the user to run and return its output. Never request or use `sudo`, root access, or privilege escalation for inspection.

## Inputs

Use the detailed intake under `Resolve the Request`.

- **Required:** API surface, target type, administrator and target-scope authority, installation scope and prefix, launcher/bootstrap selection or `recommend`, and enough target-system facts to assess documented requirements.
- **Optional:** A complete NVSHMEM version, defaulting to `latest`, and a preferred installation method.
- **Conditional:** Source-build configuration when source is selected.
- **Probe choice:** Required before direct target inspection: user-run probe or agent-run probe.
- **Execution approval:** Optional and not approved by default; approval applies only to the exact command block presented.

Use the first non-conflicting value in this source order:

1. State file.
2. Explicit prompt arguments.
3. Agent context.
4. Free-form user prompt.

Report conflicts instead of silently merging them. Execution always requires current explicit approval for the exact command block presented; a missing response is not approval.

## Limitations

- Depend on live official documentation for exact version requirements and commands. If retrieval fails, provide mapped official links and identify unverified claims instead of synthesizing an answer.
- Treat the system probe as a snapshot of the current shell and node; it cannot establish cluster-wide consistency, site policy, container-host configuration, or an untested network path.

## Instructions

### Safety and Defaults

- Default to providing commands, not executing them. An install request expresses intent, not approval.
- Never request or use `sudo`, root access, or privilege escalation for inspection. Use only an existing unprivileged shell; otherwise give the user the probe to run.
- Treat downloads, repository changes, package operations, container pulls/builds, source builds, and install-prefix writes as installation actions. Before executing any such action, present the exact command block, effects, and paths, then obtain explicit approval for that block. Silence is not approval; revised commands require fresh approval.
- Do not change drivers, kernel modules, system configuration, modulefiles, or shared cluster state unless separately requested and approved.
- Stop on the first failed install command and preserve its output. Permit at most three separately approved revised attempts before handing off.

### Resolve the Request

Before running any target probe or other system-inspection command, ask for and resolve the installation intake. Do not infer an answer from the current shell when the user has not yet chosen the intended target or installation scope.

Ask the following questions verbatim in one numbered message. Omit only a question whose answer is already explicit, unambiguous, and non-conflicting in the request or state file. Do not probe, retrieve version-specific documentation, or propose an installation method until the user answers Question 8.

1. **API:** Which NVSHMEM interface do you need: `C/C++`, `Python (NVSHMEM4Py)`, or `both`?
2. **Version:** Which complete NVSHMEM version do you want? Reply `latest` to use the newest fully documented release.
3. **Target:** Where should NVSHMEM run: `HPC cluster / multi-node system` or `single machine`?
4. **Authority:** Do you have administrator rights **and** authority to install system-wide software in the target scope? Reply `administrator` or `unprivileged`.
5. **Scope:** Where should NVSHMEM be installed: `user-owned prefix`, `system packages`, `container`, `virtual environment`, ? Include the exact prefix or environment path when it is not a container or system packages.
6. **Method:** Do you prefer a method: `binary archive`, `system repository/package`,  `source build`, `HPC SDK container`, `PyPI`, `Conda`,  ? Reply `recommend` if you have no preference.
7. **Probe:** May I run the bundled read-only system probe on the target shell? Reply `agent-run`, or reply `user-run` and I will give you the command to run and paste back.
8. **Launcher:** Which launcher/bootstrap should the application and two-PE validation use: `mpirun/mpiexec`, `nvshmrun`, `srun`, `oshrun`, another launcher, or `recommend`?

Interpret `administrator`, `unprivileged` exactly as stated. If Question 4 is unanswered after one prompt, record `unknown` and apply the unprivileged planning default from `Safety and Defaults`. Reject a scope that the recorded authority cannot modify and offer feasible alternatives. If Question 2 supplies an incomplete version such as `major.minor`, treat it as a prefix query: use `$nvshmem-docs` to resolve the newest fully documented matching release, show the complete resolved version, and obtain confirmation before executing version-specific commands. Otherwise preserve the literal requested version or `latest`. Treat Question 8 as launcher/bootstrap intent, not proof that the launcher is installed or compatible. After the probe, verify an explicit selection against the observed commands and versioned documentation. For `recommend`, select the compatible launcher implied by the execution environment when unambiguous; ask when multiple compatible choices remain. If none is available, record `none` rather than inventing a launcher, continue install planning when a launcher is not an installation prerequisite, and mark multi-PE runtime validation as blocked until a compatible launcher is available.

After Question 8 is answered, run or provide the bundled probe as selected. Use the chosen unprivileged probe for every fact it can collect, and request only the remaining facts. Ask the following conditional questions only when their condition applies:

- **Source build:** “What GPU and network topology will the build support (nodes, GPUs per node, and interconnect)? Which RDMA NICs/link layer or fabric provider are required? Which CUDA architectures must be built? Should the build include Python bindings, tests, examples, Hydra, or packages?” Read [source-build.md](references/source-build.md).
- **Managed Slurm/HPC cluster:** “May the resulting install be visible from every intended compute node, and does site policy permit the selected user prefix or container runtime?”
- **Shared cluster prefix:** “Who owns this prefix, and do all participating compute nodes mount the same path?”
- **Existing or new container:** “Which runtime will be used (`Docker`, `Apptainer/Singularity`, or another approved runtime), and is GPU and required network-device passthrough permitted by site policy?”
- **Python or both:** “Which Python interpreter or existing environment should contain NVSHMEM4Py? May I create a new virtual environment if needed?”

If the requested API is `Python (NVSHMEM4Py)` and the requested method is `binary archive` or `system repository/package`, reject that method combination before selecting artifacts or presenting commands. NVSHMEM4Py must be installed through a supported Python package channel (`PyPI` or `Conda`); select `source build` only when the user explicitly needs to build the bindings. Do not treat a native NVSHMEM archive or system NVSHMEM package as an NVSHMEM4Py installer or request one from the user. For `both`, an archive or system package may supply the C/C++ runtime, but NVSHMEM4Py still requires its own PyPI, Conda, or explicitly requested source-build installation.

Record every answer in `Target Summary`, marking supplied answers as observed user intent and unanswered optional values as `unknown`.

Record the resolved launcher and applicable bootstrap in `Target Summary` or `Requirements Check`, using `none` or `not applicable` explicitly when appropriate. A Python import, native-library load, or single-PE run does not replace the two-PE same-node runtime test. Never silently omit that test when launcher, GPU, scheduler, or allocation prerequisites are unavailable.

After the user selects a probe path, do not ask for information that can be obtained through the chosen unprivileged inspection.

### Consult Live Documentation

Use `$nvshmem-docs` before stating requirements, compatibility, supported versions, package names, repository URLs, image tags, dependency versions, build variables, or exact commands. When working from this source tree and the sibling skill is not registered, read and follow `../nvshmem-docs/SKILL.md` directly.

- Pass the normalized complete version or the literal `latest`.
- Resolve `latest` to the newest NVSHMEM release whose version, requirements, and compatibility are established by the live NVSHMEM documentation. Treat a Python binding's PyPI/Conda version as an independent package version, not as an NVSHMEM release candidate: select its current compatible released artifact from its package metadata and report the two versions separately. Apply the separate-candidate rule only when an artifact represents a newer NVSHMEM runtime/library release than the latest matching documentation. Do not apply older-release compatibility claims to such a runtime candidate.
- Request the installation guide, release compatibility and limitations, package or source procedure, and NVSHMEM4Py page when relevant.
- For NVIDIA HPC SDK containers, also request the NGC image and container-guide routes. Keep the HPC SDK/image version separate from the bundled NVSHMEM version.
- Hardcode only the stable method families and decision process in this skill. Never assume that a package name, CMake option, numeric requirement, or container tag from a previous release is still valid.
- For binary archives and Linux package downloads (DEB, RPM, or Ubuntu packages), start from the official NVIDIA NVSHMEM downloads page: `https://developer.nvidia.com/nvshmem-downloads`. When its linked official redistribution manifest provides the exact artifact `relative_path`, resolve it against that manifest's official download base and show a copyable command for the resulting URL. Likewise, show a command when the live official page exposes a full artifact URL. Selecting `binary archive` is sufficient to present the command, but never to execute it. Include this notice with the command: "By downloading and using the software, you agree to fully comply with the terms and conditions of the NVIDIA Software License Agreement." If NVIDIA requires authentication or interactive license acceptance and neither an exact URL nor an official manifest path is available to the current session, report that external limitation and provide the official download page. Never construct an artifact filename or URL from a remembered naming pattern.
- For every source build, obtain the selected release from the official NVIDIA NVSHMEM GitHub releases page: `https://github.com/NVIDIA/nvshmem/releases`. Use the release's published source archive or release tag, verify that its version matches the resolved NVSHMEM version, and report the release URL as provenance.
- If live documentation cannot be retrieved, return the official routed links and identify the blocked claims. Do not invent installation commands from memory.
- If official pages conflict, report the conflict and avoid execution until it is resolved.

### Inspect and Assess the Target

Only after the installation intake is complete and the user explicitly asks the agent to do so, run the bundled read-only probe without arguments:

```text
run_script("scripts/collect-system-info.sh")
```

If the user elects to run it themselves, provide that command and wait for its output. Keep the probe unprivileged. On a cluster, inspect representative compute nodes only within an existing allocation; otherwise provide the script or equivalent commands for the user to run there.

Compare the collected facts with the exact-version documentation and classify each relevant requirement:

- `pass`: the target satisfies the documented requirement.
- `fail`: the target contradicts the documented requirement. Stop and provide remediation choices.
- `unknown`: the available evidence is insufficient. Ask for the smallest missing fact and do not execute installation commands.
- `not applicable`: the requirement belongs only to an unselected optional feature.

For a missing optional transport dependency, do not enable that transport and explain the capability loss. If it is the NIC-matched baseline remote transport, ask before proceeding without remote transport support. Do not weaken a mandatory requirement.

When several features or transports are enabled, require the union of their documented prerequisites. Never disable a documented prerequisite while leaving the dependent feature enabled.

For every source build, select a remote transport before generating the configuration. When the probe finds a usable RDMA NIC or a supported fabric provider, choose the matching baseline remote transport from [source-build.md](references/source-build.md) and assess its prerequisites. Do not omit remote support merely because the first validation is same-node.

Only a `pass` verdict permits a transport to be enabled by default. An `unknown` prerequisite blocks generation of an exact source configuration until the missing fact is resolved; do not enable the transport or silently omit it. A `fail` verdict disables that transport and requires the capability loss to be reported, with user acceptance when it removes the NIC-matched baseline remote path.

When no remote transport is selected—because the user limits the scope to local GPUs, no matching NIC/provider is present, or the user accepts a documented prerequisite failure—leave every remote-transport support option off. If a matching transport has an unknown prerequisite, obtain the missing fact or ask the user before generating the installation command.

### Choose, Execute, and Validate

After the requirements check, read [installation-matrix.md](references/installation-matrix.md) to select a method. For a source build, also read [source-build.md](references/source-build.md). Before presenting commands, requesting execution approval, or validating an installation, read [execution-and-validation.md](references/execution-and-validation.md). On any planning, installation, or validation failure, read [failure-handoff.md](references/failure-handoff.md).

## Troubleshooting

For any planning, installation, or validation failure, preserve the command, output, target summary, selected transport/bootstrap, and documentation version, then follow [failure-handoff.md](references/failure-handoff.md). For post-install launch, initialization, runtime, or validation failures, use `$nvshmem-troubleshoot-and-report-bugs` when available; otherwise provide official support guidance with the preserved evidence.

## Available Scripts

| Script | Purpose | Arguments |
| --- | --- | --- |
| [`scripts/collect-system-info.sh`](scripts/collect-system-info.sh) | Collect read-only, unprivileged system, GPU, CUDA, launcher, RDMA, and package-manager information for installation planning. | None |

## Examples

Request: “Install the latest NVSHMEM C/C++ development package on this Ubuntu system.”

Resolve `latest`, ask about administrator authority, inspect the system, and recommend APT only when the user confirms authority over the target. Present exact commands and validation before requesting execution approval.

Request: “Build NVSHMEM from source with IBGDA for these Infiniband-connected GPUs.”

Inspect the GPU, NIC, launcher, and dependency environment; verify the requested release and IBGDA prerequisites through `$nvshmem-docs`; then present a complete CMake configuration, installation sequence, and same-node and two-node validation plan before requesting approval.

<!-- ci trigger -->

© NVIDIA NVSHMEM Team · Apache-2.0 · NVIDIA/skills

Похожие скиллы

Все скиллы по теме «DevOps и инфраструктура» →
huawei-cloud-install-openjiuwenswarm huaweicloud Скилл huawei-cloud-install-openjiuwenswarm для Claude Code автоматически загружает, настраивает и запускает JiuwenSwarm внутри контейнера разработки Huawei Cloud без ручного ввода команд. CX Codex +1 153установок runpod-mcp runpod Управляет инфраструктурой Runpod через вызовы инструментов подключённого MCP-сервера. Направляет команды на создание подов, эндпоинтов и проверку каталога GPU. CX Codex +1 14 500установок pod-doctor runpod Скилл pod-doctor находит причины сбоев в подах Runpod: ошибки 502, пропажу GPU после перезагрузки или зависание при старте. Диагностирует проблемы только в режиме чтения. CX Codex +1 3 700установок spark-training-gotchas Seth Hobson Скилл spark-training-gotchas диагностирует 10 частых сбоев при обучении моделей на NVIDIA DGX Spark. Помогает найти утечки памяти, троттлинг и ошибки CUDA перед долгими запусками. CX Codex +1 2 500установок system-profile wanshuiyin Скилл system-profile профилирует скрипты, процессы, память и GPU. Находит узкие места в коде и готовит отчёт с замерами. CX Codex +1 527установок serverless-modal wanshuiyin Скилл serverless-modal помогает запускать обучение и инференс на GPU в облаке Modal без ручной настройки SSH и Docker. Оценивает расходы и генерирует лаунчеры прямо в чате. CX Codex +1 515установок

Из того же набора312

Весь набор →
nvshmem-collect-performance-data Скилл nvshmem-collect-performance-data собирает замеры пропускной способности и задержки NVSHMEM put/get вместе с топологией узлов для проверки производительности GPU-кластера. CX Codex +1 ★ 4 106звёзд репо doca-erasure-coding Скилл doca-erasure-coding помогает работать с Reed–Solomon-кодированием через DOCA: создавать контрольные блоки, восстанавливать утраченные и обновлять чётность после изменения данных. CX Codex +1 ★ 4 106звёзд репо doca-spcx-cc Скилл помогает запускать и проверять алгоритм управления перегрузкой SPC-X на BlueField с DPA-процессором через CLI doca_spcx_cc, включая выбор между SPCX и doca-pcc. CX Codex +1 ★ 4 106звёзд репо doca-firefly Помогает настроить контейнер DOCA Firefly на NVIDIA BlueField для точной синхронизации времени по PTP. CX Codex +1 ★ 4 106звёзд репо tao-run-deft-od-aoi Скилл tao-run-deft-od-aoi для Claude Code автоматизирует итеративное обучение детектора дефектов RT-DETR на промышленных линиях. Измеряет AP50, находит похожие на ошибки образцы через SigLIP и выбирает лучшую модель. CX Codex +1 ★ 4 106звёзд репо rtvi-cv-customize-model Скилл rtvi-cv-customize-model заменяет базовый детектор DeepStream на собственную ONNX-модель в пайплайне VSS Alerts Blueprint в режиме проверки 2d_cv. CX Codex +1 ★ 4 106звёзд репо cdf-video-data-augmentation Скилл запускает пайплайн CDF Video Data Augmentation через Airflow: создаёт варианты видео в Cosmos Transfer и обрабатывает исходные и прошедшие проверку ролики. CX Codex +1 ★ 4 106звёзд репо doca-rmax Скилл doca-rmax для отладки приёма точно синхронизированных медиапотоков (SMPTE ST 2110, данные маркетов, научные каналы) на BlueField DPU через DOCA Rivermax. Помогает, когда пакеты не приходят, растёт джиттер или кривая лицензия. CX Codex +1 ★ 4 106звёзд репо

Спонсоры проекта

Компании, которые помогают держать аналитику открытой