Impossible Rooftop Leap · skill pública de Open Skill

Impossible Rooftop Leap

Impossible Rooftop Leap es una Open Skill pública en Digen. Ábrela para ver qué crea este flujo.

@HarperDiaz

Sports

6 0 6

https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4284956/2026-10-17/64evzqk3.webp

# Impossible Rooftop Leap

No glow, no effects, just a jump that shouldn't be possible. You leap a gap between skyscrapers and stick the landing — shot like real handheld phone footage.

## When to use

- The user wants a video of themselves pulling off an impossible rooftop stunt that looks like real phone footage (impossible rooftop leap, unbelievable parkour jump, no-effects stunt clip)
- Another take with a different length or orientation

Not for: Still images only.

## Inputs and gates

| Input | Required | Default | Gate |
|---|---|---|---|
| Photo | Yes | — | One adult, face clear. Ask for another photo if the face is tiny, blurred or hidden |
| Description | No | tuned default below | What the user wants; merged into the prompt |
| Duration | No | `6 s` | 6 s (default), or 10 s |
| Mood | No | `hyped disbelief` | hyped disbelief (default), calm and cool, comedic panic, or user-described |
| Stunt | No | `rooftop gap leap` | rooftop gap leap (default), wall-run flip, or user-described |
| Orientation | No | `9:16` portrait | Other ratios on request |

- If the person may be under 18, stop and decline politely. Real-person photos must be of adults the user has the right to use.
- Video does not allow NSFW content: keep wardrobe fully covering and refuse sexual requests. Suggestive fashion looks stay as still images.

## Steps

1. Read `references/video-motion.md`.
2. **Identity read**: `recognize_image(image_url=<photo>, prompt='Describe this adult for identity preservation: face shape, eye shape and color, brows, nose, lips, skin tone, apparent age range, hair color and length, distinctive marks, glasses. One compact line.')`.
3. **Keyframe**: `generate_image(prompt=<KEYFRAME>, input_urls=[photo], model='i2i.hd', orientation='portrait', aspect_ratio='9:16', name='Impossible Rooftop Leap keyframe', entity_type='work_in_progress')`. If the face drifts, regenerate once.
4. **Video**: `generate_video(prompt=<MOTION>, input_urls=[keyframe], model='i2v', orientation='portrait', aspect_ratio='9:16', duration=<d>, name='Impossible Rooftop Leap', entity_type='final_videos')`. Pass only the keyframe, never the original upload.
5. Reply in one or two sentences, e.g. "All set. Your impossible rooftop leap is on the canvas. Next I can do 10 s." Results appear on the canvas; never paste URLs.

## Prompt templates

**KEYFRAME**

```text
The same adult from the reference: <identity line>. Same face structure, eyes, skin tone and recognizable likeness; no beautification or face reshaping. Opening frame of this scene: the same adult on a sunny high-rise rooftop edge glancing at the camera in disbelief, then sprinting and launching into an impossible leap across a wide gap between skyscrapers, soaring with arms out against a clear blue sky above the city, and landing in a crouch on the opposite rooftop with a triumphant shout; shot like real handheld phone footage with natural daylight, slight motion blur and true-to-life physics, fully clothed casual streetwear, no glow or visual effects. Posed for motion, clear subject, cinematic light. Nothing from the original photo's background remains.
```

**MOTION** (motion only; the frame already shows the look)

```text
Starting from this exact frame: <one clear motion arc for this concept: the same adult on a sunny high-rise rooftop edge glancing at the camera in disbelief, then sprinting and launching into an impossible leap across a wide gap between skyscrapers, soaring with arms out against a clear blue sky above the city, and landing in a crouch on the opposite rooftop with a triumphant shout; shot like real handheld phone footage with natural daylight, slight motion blur and true-to-life physics, fully clothed casual streetwear, no glow or visual effects>, beginning, peak and settle in one continuous shot. Camera: <one move: slow push-in, gentle orbit, or static>. The same face and outfit throughout. Cinematic, no text.
```

## Failure handling

- `success=false`: say what failed in one s