# cdf-event-video-generation

cdf-event-video-generation — a skill for Claude Code and Codex that runs a DAG on Kubernetes or OSMO to transform images into anomaly videos, auto-label them, and build datasets.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf (Devin Desktop), Claude.ai · ★ 4106 repo stars
- Author: NVIDIA
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/NVIDIA/skills/tree/HEAD/skills/cdf-event-video-generation/SKILL.md
- Updated: 2026-10-23
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/skills/cdf-event-video-generation

## What it does

This skill orchestrates the full pipeline: image preparation, video synthesis via Cosmos3, automated labeling (detection, tracking, captioning, visual QA), dataset generation, and result retrieval. Routes between the Kubernetes DAG (`event_video_generation_dag_k8s`), which can use internal services, and the OSMO DAG (`event_video_generation_dag_osmo`), which requires external endpoints only.

For OSMO DAG, set `external_services: true` and supply URLs for external services (VLM, LLM, image2video, optionally Cosmos Evaluator). Before triggering OSMO, verify `OSMO_TOKEN`, `OSMO_URL`, and `OSMO_POOL` in your cluster — they may be empty even if the controller is healthy, and failures appear only when a task reaches OSMO. Defer augmentation questions to the `cdf-augmentation` skill and auto-labeling questions to `cdf-auto-labeling`.

## Requirements
- Works with: Kubernetes

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add NVIDIA/skills --skill cdf-event-video-generation -a claude-code
```

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

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add NVIDIA/skills --skill cdf-event-video-generation -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/cdf-event-video-generation/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add NVIDIA/skills --skill cdf-event-video-generation -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/cdf-event-video-generation/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install NVIDIA/skills skills/cdf-event-video-generation
```

2. **Manually: put the files in place**
- File: `.github/skills/cdf-event-video-generation/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add NVIDIA/skills --skill cdf-event-video-generation -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/NVIDIA/skills --path skills/cdf-event-video-generation
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/cdf-event-video-generation/SKILL.md` — The legacy .windsurf/skills/ path still works

### 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/cdf-event-video-generation/SKILL.md

## When to use it
- generate event and anomaly videos from a seed image
- auto-label videos with detection, tracking, and descriptions
- build datasets for safety and surveillance model training
- create video clips of falls, fights, fires, shoplifting from a template image

## Example prompts
- Run the Event Video Generation DAG on Kubernetes to synthesize a person-falling video from a photograph
- Generate a video dataset with fights and specify an external VLM service URL
- Trigger image-to-video synthesis with auto-labeling and quality checks via Cosmos Evaluator

## Similar
- [physical-ai-infrastructure-setup-and-resilient-scaling](https://affpapa.org/en/nevibecode/skills/physical-ai-infrastructure-setup-and-resilient-scaling.md) — skill · ★ 8052 repo stars
- [gke-manifest-generation](https://affpapa.org/en/nevibecode/skills/gke-manifest-generation.md) — skill · 4600 installs
- [huawei-cloud-cce-kubernetes-event-analyzer](https://affpapa.org/en/nevibecode/skills/huawei-cloud-cce-kubernetes-event-analyzer.md) — skill · 162 installs
- [qianwen-video-generation](https://affpapa.org/en/nevibecode/skills/qianwen-video-generation.md) — skill · 6100 installs
- [video-generation](https://affpapa.org/en/nevibecode/skills/video-generation.md) — skill · 3000 installs
- [eachlabs-video-generation](https://affpapa.org/en/nevibecode/skills/eachlabs-video-generation.md) — skill · 525 installs

From the same set: [nvidia/skills](https://affpapa.org/en/nevibecode/by/nvidia/skills.md) — 312 cards

---
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
