cdf-orchestration-setup — Claude Code skill: deploy Airflow controller
cdf-orchestration-setup — a skill for end-to-end deployment and connection of CDF Orchestration Airflow controller on Kubernetes. Runs readiness check, builds image, installs via Helm, and verifies connectivity.
no license ★ 4,106 updated 2026-10-23
npx skills add NVIDIA/skills --skill cdf-orchestration-setup -a claude-codeWhat cdf-orchestration-setup does
This skill handles the full deployment sequence for CDF Orchestration Airflow controller after preflight validation confirms the environment is ready. It runs make setup to initialize secrets, builds the controller image with DAGs and plugins baked in (or uses a prebuilt image), installs or upgrades the Helm release, and runs post-install connectivity and port-forwarding checks.
Before each change, the skill presents exact commands and their consequences, requiring explicit user approval. Deployment fully clears the target namespace: cronjobs, deployments, services, pods, and PVCs (including model cache and logs) are removed. DAGs and plugins are embedded in the controller image and not synced from S3 at runtime. OSMO workflows require external endpoints; in-cluster model services work only in Kubernetes mode.
The skill does not repeat prerequisite checks, does not support K3s, and does not cover DAG authoring, payload execution, or standalone diagnostics. It preserves parameters and findings from preflight verification throughout the deployment.
What it helps with
- deploy CDF Orchestration Airflow controller on a Kubernetes cluster
- upgrade the controller image and reinstall the Helm release
- connect external endpoints for models and services
- verify controller connectivity and port-forward access
- prepare a cluster and environment for OSMO workflows
Example prompts for the agent
- Deploy the CDF Orchestration controller using the image from
make/common.mk - Reinstall the controller with a prebuilt image
registry.example.com:v1.0 - Check if the cluster is ready for deployment and show the install commands
How to install cdf-orchestration-setup
1Skills CLI (Vercel)
npx skills add NVIDIA/skills --skill cdf-orchestration-setup -a claude-code2Manually: put the files in place
~/.claude/skills/cdf-orchestration-setup/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/cdf-orchestration-setup/
1Skills CLI (Vercel)
npx skills add NVIDIA/skills --skill cdf-orchestration-setup -a codex2Manually: put the files in place
.agents/skills/cdf-orchestration-setup/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add NVIDIA/skills --skill cdf-orchestration-setup -a cursor2Manually: put the files in place
.cursor/skills/cdf-orchestration-setup/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install NVIDIA/skills skills/cdf-orchestration-setup2Manually: put the files in place
.github/skills/cdf-orchestration-setup/SKILL.md1Skills CLI (Vercel)
npx skills add NVIDIA/skills --skill cdf-orchestration-setup -a gemini-cli2Platform CLI
gemini skills install https://github.com/NVIDIA/skills --path skills/cdf-orchestration-setupManually: put the files in place
.devin/skills/cdf-orchestration-setup/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about cdf-orchestration-setup
How do I install cdf-orchestration-setup in Claude Code?
Run npx skills add NVIDIA/skills --skill cdf-orchestration-setup -a claude-code, then open ~/.claude/skills/cdf-orchestration-setup/SKILL.md. For Codex, use npx skills add NVIDIA/skills --skill cdf-orchestration-setup -a codex.
What does cdf-orchestration-setup do?
It performs end-to-end CDF Orchestration controller deployment: verifies environment readiness, builds the Airflow image or uses a prebuilt one, installs the Helm release, and checks connectivity. Every operation requires explicit user confirmation.
What happens when the controller is reinstalled?
Deployment fully clears the target namespace, removing cronjobs, deployments, services, pods, and potentially PVCs (including model cache and logs). Back up critical data before reinstalling.