cdf-video-data-augmentation — Claude Code skill: what it does
This skill runs the CDF Video Data Augmentation workflow through Airflow, creating Cosmos Transfer video variants and processing source and passing generated videos.
no license ★ 4,106 updated 2026-10-23
npx skills add NVIDIA/skills --skill cdf-video-data-augmentation -a claude-codeWhat cdf-video-data-augmentation does
The skill prepares a payload for an existing DAG, checks the environment, starts the workflow through Airflow, and tracks its run. The pipeline processes source videos through auto-labeling and training export, creates Cosmos Transfer variants, then processes variants that pass checks and validates storage outputs. It does not create a single manifest for the final dataset.
Choose Kubernetes or OSMO as the execution target. Kubernetes is the default; OSMO needs external VLM, LLM, Cosmos Transfer, and required Cosmos Evaluator endpoints, with external_services: true and every service_lifecycle.*.enabled value disabled. OSMO cannot deploy those services, so endpoint setup is a prerequisite. Container configuration for augmentation and auto-labeling belongs to separate component skills, while structural DAG changes are outside this skill’s scope.
What it helps with
- prepare a payload for an existing VDA DAG
- check Airflow and the execution environment before a run
- start the workflow and monitor its status
- verify and retrieve video-processing artifacts
Example prompts for the agent
- Prepare and validate a payload for the existing VDA DAG on Kubernetes.
- Run the Video Data Augmentation workflow through Airflow and report its status.
- Check the workflow outputs and help me retrieve the generated artifacts.
How to install cdf-video-data-augmentation
1Skills CLI (Vercel)
npx skills add NVIDIA/skills --skill cdf-video-data-augmentation -a claude-code2Manually: put the files in place
~/.claude/skills/cdf-video-data-augmentation/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/cdf-video-data-augmentation/
1Skills CLI (Vercel)
npx skills add NVIDIA/skills --skill cdf-video-data-augmentation -a codex2Manually: put the files in place
.agents/skills/cdf-video-data-augmentation/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add NVIDIA/skills --skill cdf-video-data-augmentation -a cursor2Manually: put the files in place
.cursor/skills/cdf-video-data-augmentation/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install NVIDIA/skills skills/cdf-video-data-augmentation2Manually: put the files in place
.github/skills/cdf-video-data-augmentation/SKILL.md1Skills CLI (Vercel)
npx skills add NVIDIA/skills --skill cdf-video-data-augmentation -a gemini-cli2Platform CLI
gemini skills install https://github.com/NVIDIA/skills --path skills/cdf-video-data-augmentationManually: put the files in place
.devin/skills/cdf-video-data-augmentation/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-video-data-augmentation
How do I install cdf-video-data-augmentation in Claude Code?
Run npx skills add NVIDIA/skills --skill cdf-video-data-augmentation -a claude-code. The skill file is at ~/.claude/skills/cdf-video-data-augmentation/SKILL.md.
What does cdf-video-data-augmentation do?
It runs the existing Video Data Augmentation workflow through Airflow. The workflow creates Cosmos Transfer variants, processes source and passing generated videos, and checks storage outputs.
What does an OSMO run require?
Provide external VLM, LLM, Cosmos Transfer, and required Cosmos Evaluator endpoints. Set external_services: true and disable every service_lifecycle.*.enabled value, since OSMO cannot deploy these services.