skills.rest
NVIDIA
NVIDIAOFFICIAL
Explore all skills by this author →
WebsiteGitHub

holohub-app-lifecycle

Guides scaffolding, building, testing, and benchmarking HoloHub applications through the ./holohub CLI workflow.

3.2k|370|Updated Feb 25, 2026
Search Tags:#benchmarking#Testing#Scaffolding#application lifecycle#container build#holoscan#holohub
One-click install
npx skills add https://github.com/NVIDIA/skills --skill holohub-app-lifecycle
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: holohub-app-lifecycle
Source: https://github.com/NVIDIA/skills/tree/main/skills/holohub-app-lifecycle
Command: npx skills add https://github.com/NVIDIA/skills --skill holohub-app-lifecycle

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Building HoloHub applications involves many error-prone steps: resolving the right checkout, scaffolding new apps, running containerized builds, producing finite test evidence, and benchmarking data flows. This Skill gives AI agents a safe, repeatable procedure for the entire non-failing HoloHub application lifecycle using the public ./holohub wrapper.

Core Features & Use Cases

  • Checkout and workspace safety: Resolves one valid HoloHub checkout, preserves dirty state and unrelated work, and requires authorization before cloning or mutating anything.
  • Preview-then-act workflow: Every effect-bearing ./holohub command (setup, create, build, run, test) is previewed with --dryrun before the identical real command executes.
  • Finite evidence and benchmarking: Enforces bounded run modes, visual/recording inspection, lint, and optional flow benchmarking with raw-log validation.
  • Use Case: Ask your agent to scaffold a new Python HoloHub app that replays a video with a Holoviz overlay, add a 30-frame headless smoke test, and produce a reproducible flow benchmark — the Skill walks through setup preview, create, build, run, test, and benchmark in order.

Quick Start

Ask your agent to scaffold, build, run, and test a new HoloHub application from your brief using the ./holohub workflow.

Frequently Asked Questions about holohub-app-lifecycle

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I scaffold a new HoloHub application with ./holohub?▼

First preview template setup with ./holohub setup --scripts template --dryrun, get authorization for the host dependency install, run the real setup, then preview and run ./holohub create <app> --language <cpp-or-python> -i False. Stop before create if setup fails.

How do I build and run a HoloHub app safely?▼

Preview each command with --dryrun --verbose, then run the identical command without it, keeping project, mode, language, and inputs identical. Use the container-first path and verify a finite verdict or visual evidence rather than relying on exit code zero alone.

When should I use holohub-app-lifecycle versus holohub-debug-build-run?▼

Use holohub-app-lifecycle only for non-failing application work such as scaffolding, building, testing, and benchmarking. Route any concrete failing ./holohub command to holohub-debug-build-run, preserving the exact command, revision, and dirty state.

Does the skill work with a dirty HoloHub checkout?▼

Yes, but only when the user explicitly selected that checkout and the requested paths provably do not overlap existing changes. Auto-discovered checkouts must be clean, and branch switches in a dirty checkout require user authorization.

When is it safe to use --no-docker-build or --no-local-build?▼

Use --no-docker-build only after one matching build/run proves the image is unchanged. Use --no-local-build only when current artifacts or mounted-source execution are proved sufficient, and rebuild after any image or setup change.

What are the limitations of the HoloHub app lifecycle workflow?▼

It never runs sudo ./holohub, recursively scans the home directory, overwrites nonempty destinations, or commits without request. It also does not support module DEB/WHEEL packaging (use holohub-module-lifecycle) or first-time SDK host installation (use holoscan-setup).

Related Skills

Explore More Software Engineering Skills→
sayalinvidiasayalinvidia

holoscan-install-container

Install Holoscan SDK in a Docker container and validate GPU-enabled execution.

Community
Advanced
dmae97dmae97
130

omk-test-debug-loop

Classify failing command output and generate root-cause fix plans.

Community
Intermediate
sayalinvidiasayalinvidia

holoscan-install-debian

Install the Holoscan SDK on Ubuntu via apt for C++ development.

Community
Advanced
sayalinvidiasayalinvidia

holoscan-install-conda

Installs Holoscan SDK and dependencies into a Conda CUDA 13 environment.

Community
Intermediate