菜单
Generate UGC-style video prompts for Seedance 2.0 on Higgsfield for Beauty, Apparel, and Home & Lifestyle brands. Use when creating influencer-style product reviews, demos, unboxings, lipsync, or lifestyle ads. Based on Sirio Berati's AI UGC workflow. Source: sirioberati/Seedance-2.0-AI-UGC
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
Cinematic Seedance 2.0 video prompt skill for Food & Beverage brands. Use for CPG, snacks, drinks, supplements, sauces, meal kits, any edible product. Produces Higgsfield-ready prompts with hook, money shot, sound, and lighting specs. Source: beshuaxian/higgsfield-seedance2-jineng skill 14
One-question-at-a-time client interview skill. Run before any content generation. Drills down on brand, customer, product, taste, and goals. Produces brief.yaml. Based on mattpocock/skills grill-me pattern. Never batch questions.
Score social media content units for viral potential (0-100) with transparent breakdown by dimension. Use after content generation, before surfacing to human for approval. Produces ranked output with reasoning. Required for all content.
Anti-AI writing guardrails for all verbal content. Use on every caption, hook, script, email subject, button label, and UI copy before it leaves the pipeline. Triggers automatically on all text output. Source: hardikpandya/stop-slop
基于 SOC 职业分类
| name | ugc-video |
| description | Generate UGC-style video prompts for Seedance 2.0 on Higgsfield for Beauty, Apparel, and Home & Lifestyle brands. Use when creating influencer-style product reviews, demos, unboxings, lipsync, or lifestyle ads. Based on Sirio Berati's AI UGC workflow. Source: sirioberati/Seedance-2.0-AI-UGC |
| source | https://github.com/sirioberati/Seedance-2.0-AI-UGC |
Seedance 2.0 UGC uses two fundamentally different prompt formats:
Write exactly what happens, frame by frame. Who, what, how, where, energy. Use when: you have a specific vision. Max control.
Describe the result you want, not the steps to get there. Use when: you want creative variation. Less control, more surprise.
Default to Style A for production ads. Use Style B for creative exploration.
Using @image1 as the product reference, create a [UGC style descriptor] video
featuring a [specific person description — age, energy, style, NOT ethnicity].
They are [specific action with product — what exactly are they doing?] while
[environment description — where, lighting, what's around them].
They [specific gesture or movement] as they [speak to camera / hold product /
demonstrate use]. The composition is [vertical 9:16 / horizontal 16:9].
Lighting is [soft natural / warm studio / golden hour / ring-lit].
Energy is [casual recommendation / excited discovery / expert review /
aspirational lifestyle / ASMR calm / before-after reveal].
The tone feels like [specific comparison — a friend's story, a Reddit comment,
a skincare subreddit post, a fashion haul TikTok].
[Optional: add specific dialogue direction if lipsync]
Mindset: Trust, transformation, specificity, skin close-ups.
Proven formats:
Example prompt (Style A, skincare):
Using @image1 as the product reference, create a natural UGC skincare review video
featuring a woman in her early 30s with a fresh, no-makeup look and warm energy.
She is applying the serum to her face in front of a bathroom mirror, using her
fingertips to press it into her cheekbones and under-eye area. The bathroom is
softly lit with natural daylight from a window off-frame left.
She glances at camera mid-application with a slight smile — not performed, just
caught in a real moment. Composition is vertical 9:16.
Lighting is soft diffused natural. Energy is honest recommendation — like a friend
who genuinely uses this and wants you to know about it. No ring light glow.
No branded backdrop. This looks like someone's real bathroom on a Tuesday morning.
Things that kill beauty UGC:
Mindset: Aspiration, lifestyle, how it moves, how it makes you feel.
Proven formats:
Example prompt (Style A, apparel):
Using @image1 as the product reference, create a natural try-on style TikTok
featuring a woman in her mid-20s with a casual, confident style — jeans and
white tee as baseline, switching into the featured piece.
She's in a bright bedroom with good natural light, a full-length mirror visible
at frame edge. She pulls the item from a hanger, holds it up to herself,
then cuts to wearing it, doing a natural slow 360 turn.
Her hands adjust the fit at the waist — a real gesture, not staged.
She looks at herself in the mirror with a "yeah, that works" expression,
not the camera. Vertical 9:16.
Energy: the way someone actually acts when they find something that fits
perfectly. Understated satisfaction. Not a sales pitch.
Mindset: Transformation, cozy, "I made this space better", process reveal.
Proven formats:
Example prompt (Style A, home):
Using @image1 as the product reference, create a slow-reveal home styling video
showing a living room corner being transformed.
Start wide: the corner is bare except for a floor lamp. A pair of hands enters
frame and sets the featured product (a ceramic vase) on a small side table.
Then a single stem of dried flowers. Then a small stack of two books.
Each placement is deliberate — a 2-second pause between each.
Final shot: slow zoom out from the vase, revealing the full styled corner.
Warm afternoon light from window right. No voiceover. No text overlay needed —
the styling tells the story.
Composition shifts from 16:9 wide to a final vertical 9:16 close-up of the vase.
Energy: the satisfaction of getting a space exactly right. Quiet. Intentional.
For lipsync ads where you supply the voiceover:
Using @audio1 as the voiceover, create a lipsync video with a [creator description].
They are speaking directly to camera in a [setting]. Their expression matches
the energy of the audio — [calm and conversational / excited discovery / honest review].
Vertical 9:16. [Lighting description].
For Shopify apps, digital products, or software UI:
Using @image1 as the UI screenshot reference, create an iPhone-style vertical
UGC app review featuring a [creator description] in the foreground holding a
smartphone upright in one hand, clearly showing the screen to the viewer.
The phone is angled slightly toward camera so the interface is readable.
Creator is speaking directly to camera with [energy descriptor — excited /
practical / "I just found this" energy], slightly leaning forward and presenting
the phone like evidence — like they're sharing something they actually use.
The screen displays [brief description of what should appear on screen].
Composition: vertical 9:16. Creator takes up left 2/3 of frame,
phone in right hand visible at frame right.
Run 2 variants per content idea before committing to full production:
Variant A: Style A prompt (fully descriptive) Variant B: Style B prompt (outcome-driven)
Score both on:
Pick the higher scorer. Kill the other. No sentiment.
Over-directing performance — "She smiles with genuine excitement" is direction. Seedance reads it literally and produces a smile that looks forced. Better: describe the situation that would naturally produce that expression.
Forgetting the product — UGC is about the product. Every 3–5 seconds, the product should be visible or referenced.
Too much text direction — If you need 5 lines to describe what someone says, use lipsync with your own audio instead.
Generic "influencer" description — Avoid "a beautiful young influencer." Be specific: "a woman in her late 20s, undercut hair, wearing a linen shirt, the kind of person who has opinions about coffee grinders."
Ignoring platform — TikTok UGC = fast cuts, captions baked in, trending audio. Instagram Reels = slightly more polished, slower reveal. Pinterest = aspirational lifestyle, less talking-head. Prompt differently per platform.