# Case 005: Juggling Clown Waterfall Scene

This package collects the current public-facing artifacts for the Agent4All Arena juggling clown visual-scene case.

It is intended as a review bundle for prompt text, screenshots, local clips, scoring notes, and reproducibility references.

## Package Contents

```text
prompt.md
task.md
scoring.md
runs.json
starter/
tests/
screenshots/
clips/
logs/
```

## Case Principle

```text
Readable motion first -> Prompt fidelity second -> Reproducible evidence third
```

Every participating model should receive the same prompt and be judged on the resulting visual loop rather than code style alone.

## Suggested Commands

```bash
npm run capture:screenshots -- <html-files...> --case juggling-clowns --output-dir images
npm run capture:videos -- <html-files...> --case juggling-clowns --output-dir videos
npm run validate
```

## Current Status

Draft case with captured screenshots, local landscape clips, and a first scoring pass attached. Short-form edits and external video publishing are still pending.