agent-skills
Higgsfield UGC Generator Skill
Create realistic UGC-style Seedance 2.0 prompts that look like casual phone footage: selfie testimonials, talking-head clips, product reviews, unboxings, and creator-style ads.
This skill is built around the details that usually make AI UGC feel fake. It asks for the right inputs before writing, maps your uploaded images and audio into Seedance @ tags, and explicitly controls the camera setup so a handheld selfie looks handheld, a propped phone stays still, and a friend-filmed shot moves like a real person is holding the phone.
What it does
- Generates one paste-ready prompt for Seedance 2.0 instead of a long breakdown or creative brief
- Asks for the missing creative inputs: what the person is doing, what they are holding, the setting, emotion, camera setup, and audio plan
- Tags reference assets clearly with
@image1,@image2,@audio1, and@video1so Seedance knows which upload is the character, product, environment, or audio - Applies realistic phone-camera behavior: selfie tremor, walking bob, propped-phone stillness, tripod lock-off, or friend-filmed handheld drift
- Adds specific facial and body performance: eyes, eyebrows, mouth shape, head movement, hand gestures, posture shifts, and natural pauses
- Avoids overproduced video language like cinematic lighting, speed ramps, transitions, brand cards, lens flares, and dramatic grading
Best for
Use this when you want a video that feels like a person casually filmed it on a phone:
| Use case | Example |
|---|---|
| Selfie testimonial | A creator holding the phone at arm's length and talking directly to camera |
| Product review | A creator holding a product, app, package, or prop and explaining why it matters |
| Unboxing | Two hands opening packaging while the phone is propped on a desk or counter |
| Talking-head clip | A founder, customer, or creator delivering a short opinion or explanation |
| Organic ad | A casual creator-style clip that feels native to TikTok, Reels, or Shorts |
How it works
Start with a character image, starting frame, product image, or a plain text concept. The skill will ask for anything important that is missing before it writes the final prompt.
It then decides how the phone is physically set up. If one hand is free, it can describe handheld selfie tremor. If both hands are busy unboxing or holding a product, it locks the phone on a surface with no shake or drift. If someone else is filming, it adds the slower, more intentional movement of a handheld operator.
For audio, it handles both workflows: if you upload audio, the prompt tags it and keeps the visual direction separate; if you want Seedance to generate dialogue, the skill writes the speech and delivery directly into the scene.
Members get the full resource
Unlock the complete walkthrough
- Full implementation notes, examples, checklists, and templates.
- Private Discord access for feedback, accountability, and support.
- Monthly Startup Club resource drops for solo founders building with AI.
- Downloadable Higgsfield UGC Generator Skill file.