Storm Crowd Movement Clip · Open Skill 公开技能

Storm Crowd Movement Clip

Storm Crowd Movement Clip 是 Digen 上的公开 Open Skill。打开后即可了解该工作流能制作什么。

@HazelReel

Film

2 0 2

https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4321739/2026-10-21/qr8eeupf.gif

# Storm Crowd Movement Clip

One figure, still as stone. Behind them a masked crowd snaps and leans in eerie unison — cold light, drifting smoke, a hyper-real cinematic storm.

## When to use

- The user wants a cinematic video where their photo's subject stands still while a hooded crowd behind performs synchronized sharp movements (storm crowd clip, I come like a storm video, synchronized crowd cinematics)
- 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; a three-quarter or full-body shot works best so the standing pose can read. 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 |
| Crowd | No | `hooded figures in off-white shirts` | hooded figures in off-white shirts (default), faceless masked figures, dark-suited figures |
| Mood | No | `cold steel-grey` | cold steel-grey (default), deep teal shadows, ember-warm dusk |
| Duration | No | `6 s` | 6 s (default), or 10 s |
| 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='Storm Crowd Movement Clip 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='Storm Crowd Movement Clip', entity_type='final_videos')`. Pass only the keyframe, never the original upload.
5. Reply in one or two sentences, e.g. "All set. Your storm crowd movement clip is on the canvas. Next I can do faceless masked figures." 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 subject stands centered and motionless in <outfit: black turtleneck (default), or user-described> while a dense ring of <crowd: hooded figures in off-white shirts (default), faceless masked figures, dark-suited figures> surrounds them, performing synchronized head turns, shoulder shrugs and forward leans; <mood: cold steel-grey (default), deep teal shadows, ember-warm dusk>, shallow depth-of-field bokeh, physically accurate drifting smoke, hyper-real skin and fabric, fully clothed, cinematic film still. 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: the surrounding <crowd: hooded figures in off-white shirts (default), faceless masked figures, dark-suited figures> begin to move in one eerie wave — first a slow synchronized head turn toward the centered figure, then at the peak every shoulder shrugs and the whole ring leans forward in sharp unison as smoke drifts across the frame, then the crowd settles back into stillness on the final beat while the centered subject does not move at all. One slow camera push-in from a low, steady track, easing closer across the full beat. Preserve facial features a