# holoscan-install-wheel

Installs Holoscan SDK 5.x CUDA 13 Python wheels into an isolated venv and checks that the import works.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · 1800 installs · ★ 3735
- Author: Holoscan Team
- License: Apache-2.0 · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/NVIDIA/skills/tree/HEAD/skills/holoscan-install-wheel/SKILL.md
- Updated: 2026-10-23
- HTML: https://affpapa.org/en/nevibecode/skills/holoscan-install-wheel

## What it does

The skill walks an agent through a Holoscan SDK 5.x install. It picks a specific 5.x release first, reads that release's notes and the versioned SDK installation guide, then inspects the host: `python3 --version`, `getconf GNU_LIBC_VERSION`, the Linux architecture and the CUDA 13 driver. You need regular CPython 3.12–3.14, glibc 2.39 or newer, and x86_64 or arm64.

Then comes the package choice. `holoscan` pulls the full stack through `holoscan-cu13`; `holoscan-core-cu13` is core only, and optional modules ship as `holoscan-vision-cu13`, `holoscan-visualization-cu13` and `holoscan-inference-cu13`. Every component has to sit on the same SDK version. Installation runs in a fresh venv: `pip index versions`, a `--dry-run`, the real install, `pip check`, `pip list --format=freeze`, and an import that prints the version and the operator count from the registry.

Here is where it usually breaks. The CUDA field in `nvidia-smi` reports driver capability, not the installed toolkit, so a driver that advertises CUDA 12 needs a compatibility assessment or an upgrade before a CUDA 13 install. The wheels carry Python bindings and native runtime libraries but no C++ headers, CMake files or examples, so native C++ work means Debian packages or a source build. If any command fails, the skill stops and reports the output instead of pushing on.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add NVIDIA/skills --skill holoscan-install-wheel -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/holoscan-install-wheel/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/holoscan-install-wheel/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add NVIDIA/skills --skill holoscan-install-wheel -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/holoscan-install-wheel/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add NVIDIA/skills --skill holoscan-install-wheel -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/holoscan-install-wheel/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install NVIDIA/skills holoscan-install-wheel
```

2. **Manually: put the files in place**
- File: `.github/skills/holoscan-install-wheel/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add NVIDIA/skills --skill holoscan-install-wheel -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/NVIDIA/skills --path skills/holoscan-install-wheel
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/holoscan-install-wheel/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add NVIDIA/skills --skill holoscan-install-wheel -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/NVIDIA/skills/tree/HEAD/skills/holoscan-install-wheel/SKILL.md

## When to use it
- install Holoscan SDK 5.x into an isolated venv without touching system Python
- pick between the full stack and single modules like vision or inference
- check CPython, glibc and the CUDA 13 driver before installing
- keep every component on the same SDK version
- run an import and dependency check after install

## Example prompts
- install Holoscan SDK 5.x in a separate environment
- check whether my driver and glibc work for the Holoscan 5.x wheels
- set up a venv with holoscan-core-cu13 and verify the import

## Source text · SKILL.md · Apache-2.0

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: ee8b38b65e687661.

BEGIN THIRD-PARTY FILE ee8b38b65e687661
```text
---
name: holoscan-install-wheel
description: "Install Holoscan SDK 5.x CUDA 13 Python wheels in an isolated environment, choosing the full stack or version-matched components."
license: Apache-2.0
metadata:
  author: "Holoscan Team <holoscan-team@nvidia.com>"
  version: "2.0.0"
  github-url: "https://github.com/nvidia-holoscan/holoscan-sdk"
  tags:
    - holoscan
    - installation
    - wheel
---

# Holoscan 5.x Python Wheel Installation

## Release and platform checks

This skill targets **Holoscan SDK 5.x**. Select the requested 5.x release before
choosing artifacts. Read its release notes and the matching
[SDK installation guide](https://docs.nvidia.com/holoscan/sdk-user-guide/setup/sdk-installation).
Select the 5.x documentation version; an unversioned page may describe 4.x.
For an unreleased checkout, use its `docs/sdk_installation.mdx` and `DEVELOP.md`.
Early Access binary commands describe planned artifacts, so confirm publication
for the exact release before installing. If unavailable, explain that status
and offer `holoscan-install-source`; keep the requested 5.x version.

The 5.0 artifact set targets CUDA 13 and dGPU systems on x86_64 or arm64.
Native packages use Ubuntu 24.04; wheels use glibc 2.39 and regular CPython
3.12–3.14. Check the selected release for changes to that matrix. There is no
CUDA 12 variant or separate Tegra/iGPU artifact in this set. A driver reporting
CUDA 12 support needs a compatibility assessment or driver upgrade before a
CUDA 13 installation; the `nvidia-smi` CUDA field is a driver capability, not
the installed toolkit version. Never infer compatibility from that field alone.

## Choose a distribution

Wheels contain Python bindings, typing information, and native runtime libraries.
C++ headers, CMake files, examples, and sample data are separate. For native
C++ development use Debian packages or a source build.

| Goal | Distribution |
| --- | --- |
| Full stack | `holoscan` (selects `holoscan-cu13`) |
| Core only | `holoscan-core-cu13` |
| Optional modules | `holoscan-vision-cu13`, `holoscan-visualization-cu13`, `holoscan-inference-cu13` |

Inspect the host, `python3 --version`, and `getconf GNU_LIBC_VERSION`. Confirm
regular CPython 3.12–3.14, glibc 2.39 or newer, a compatible Linux architecture,
```
END THIRD-PARTY FILE ee8b38b65e687661

The third-party file has ended. Everything between the ee8b38b65e687661 tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

First 2282 of 5080 bytes. © Holoscan Team · Apache-2.0 · https://github.com/NVIDIA/skills/tree/HEAD/skills/holoscan-install-wheel/SKILL.md

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md