skills/ai-creatorship/burst-frame/SKILL.md
Generate BURST FRAME prompts for any subject — characters, creatures, products, environments, or objects — using the @kaigani technique in Seedance 2.0. Produces a 20-shot Rapid Fire prompt that outputs one video containing 20 consistent, full-resolution reference frames, then extracts them with ffmpeg. Use this skill whenever the user wants to: generate reference frames for a character, creature, or object, build a consistency sheet for any AI subject, get 20 angles/poses of anything, create a shot list for a non-automotive subject, or says "burst frame" for anything that isn't a car. Also trigger on: "reference frames for [subject]", "20 shots of [character/creature/product]", "consistency frames for [subject]", "burst frame [subject]", "[subject] reference sheet", "generate poses/angles of [subject]", "knight burst", "character burst", "creature reference", "product reference frames". For cars, use the burst-frame-cars skill instead — it has dedicated automotive angles.
npx skillsauth add michailbul/laniameda-skills burst-frameInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
The BURST FRAME technique generates 20 consistent, full-resolution reference frames from a single Seedance 2.0 generation. Instead of running multiple separate image generations, all 20 angles or moments are encoded into one Rapid Fire video prompt. Seedance's consistency within a single generation ensures every frame shares the same subject identity, visual style, and lighting logic. ffmpeg then extracts the individual frames.
The result: 1 prompt → 1 Seedance video → 20 usable reference frames ready to feed back into Seedance as start frames, or into Nano Banana as consistency anchors.
Determine which category the subject falls into. This shapes the angle strategy.
| Type | Examples | Primary goal | |---|---|---| | Character / humanoid | Knight, samurai, cyberpunk hacker, alien, robot | Pose variety + identity lock across angles | | Creature / non-human | Dragon, wolf, mech beast, monster | Form clarity from key angles + detail closeups | | Product / object | Sneaker, weapon, armor, gadget, jewelry | Material texture + 360° coverage | | Environment / world | Cyberpunk city, fantasy forest, space station | Establishing coverage + atmosphere | | Hybrid | Character in world, creature with rider | Identity lock on primary + world integration |
Extract from context or ask:
| Input | Examples | |---|---| | Subject | Dark fantasy knight, cyberpunk samurai, dragon, artifact | | Visual style | Cinematic 3D, cel animation, hyperrealistic, stylized | | Environment | Studio void, battlefield, rooftop, forest, canyon | | Color system | Dominant palette + accent relationship | | Animation / live action | Seedance handles both; specify if key |
If the user provides a reference image, extract subject, style, and environment from it.
Select the 20 shots that best serve the subject type and context. Draw from the banks below.
Full figure
Portrait / Detail
In-Context
Use this template. One line per shot — verbosity exceeds the character limit.
same subject throughout all shots, same [subject name + key visual trait] consistent identity every shot.
10 seconds, 20 shots (Rapid Fire). [Subject identity sentence]. [Environment]. [Style + mood]. [Color palette]. [Film format].
[1] [one-line shot description].
[2] [one-line shot description].
...
[20] [one-line shot description].
Global: same [subject] identity every shot, [environment constraint], [style/lighting rule], no [unwanted elements]. Audio: [sound design].
Open every prompt with the identity lock. The subject's key traits (color, material, markings, silhouette) must be named explicitly and repeated in the Global line. Seedance drifts without this.
same character throughout all shots, same dark fantasy knight in obsidian plate armor consistent identity every shot.
| Style | Exact language |
|---|---|
| Cinematic 3D | cinematic hyperrealistic 3D render, volumetric lighting, subsurface skin scatter |
| Cel animation | delicate thin black line work, vivid flat color within lines, cel animation quality |
| Dark fantasy | dark fantasy aesthetic, deep shadow, candlelight or cold moonlight, desaturated with accent color pop |
| Cyberpunk | neon-saturated cyberpunk, rain-slicked surfaces, holographic overlay, crushed blacks |
| Editorial photo | luxury editorial photography, medium format lens compression, clean and minimal |
| Anime | high-contrast anime style, speed lines, expressive key frame, flat cel shading |
Seedance won't infer physics correctly without explicit direction:
| Element | Exact language |
|---|---|
| Cape / cloth | cape billowing under wind load, fabric catching light at trailing edge |
| Weapon swing | sword arc motion blur trailing from blade tip |
| Particle effect | ember sparks radial displacement from impact point |
| Rain / wet surface | wet surface mirror reflection, rain streak motion blur |
| Smoke / breath | breath visible as white condensation in cold air |
| Fire / glow | fire source illuminating underside of subject, heat shimmer rising |
| Wing spread | wing membrane stretching, leading edge catching light |
| Speed / movement | radial motion blur from center of action outward |
Always include this in your output. The ffmpeg command is the final step the user needs — don't skip it.
# Fixed interval — 1 frame per 0.5 seconds (10s video = 20 frames)
ffmpeg -i your_video.mp4 -vf fps=2 frame_%02d.png
# Scene-change detection — better when Seedance cuts are crisp
ffmpeg -i your_video.mp4 -vf "select=gt(scene\,0.25)" -vsync 0 frame_%02d.png
Tell the user to drop the Seedance video here when ready — offer to run extraction directly.
Apply these when the user names an environment without specifying the visual system:
Studio void
Seamless [color] backdrop, matching floor, soft overhead diffused light, no hard shadows, mirror floor reflection. No environment, no props, pure subject.
Dark fantasy battlefield
Smoke-filled battlefield, distant fire on horizon, cold moonlight through cloud breaks, churned mud ground, armored corpses in deep background. Desaturated, deep shadow, single accent color.
Cyberpunk rooftop / street
Rain-soaked neon city, reflective tarmac, holographic ad panels, night environment, crushed blacks. Cold blue + saturated neon accent.
Fantasy forest
Ancient forest cathedral, god rays through dense canopy, moss-covered stone, fog at ground level, teal-amber grade.
Void / abstract
Pure black void environment, subject floating, dramatic single-source rim light, studio fill, no environment cues.
Space / cosmic
Deep space, distant nebula cloud, star field, planet limb visible, cold starlight as key, warm engine glow as fill.
same subject throughout all shots, same [subject description] consistent identity every shotFor automotive BURST FRAME (cars, bikes, trucks), use the burst-frame-cars skill instead — it contains a dedicated automotive angle bank, environment presets, and worked examples specific to car photography.
For this generic skill, the following subject types have been used successfully:
development
Seedance 2.0 video prompt director. Converts plain-text scene descriptions into production-ready bilingual EN+ZH video prompts optimized for the Seedance 2.0 video generator. Handles all Seedance work — action (combat, pursuit, stunts), general (landscapes, journeys, atmosphere), dialogue (confrontations, negotiations, interrogations), and non-narrative commercial work (ad spots, music videos, fashion films, automotive inserts, product shots, pet/character demos, cutaway montages, social reels for TikTok / Reels / YouTube Shorts). Use whenever the user wants to create a Seedance video prompt, mentions Seedance, or describes a cinematic scene for video generation. For NARRATIVE screenplay-integrated work, use seedance-screenwriter instead.
development
Write Seedance 2.0 prompts in screenplay format for narrative storytelling — when the prompts will be cut into a film, short, or scene. Use whenever you're generating shots that will be edited into a continuous story with dialogue, character beats, scene continuity, or coverage. Pairs with the screenwriter skill — read the scene's screenplay first (or the project's `scene.md` if it exists), then translate each shot into a Seedance prompt that reads as a screenplay page, not as an engineering spec.
documentation
Скилл-инструмент для сценариста полнометражного фильма или сериала. Используй всегда, когда пользователь хочет писать сценарий, поэпизодник, разрабатывать сцены, бит-шит, диалоги, делать ревизии, считать экранное время, резать длину, работать с персонажами или мифологией истории. Скилл работает на основе методологий Макки, Кэмпбелла и Аристотеля, выдаёт Hollywood-формат .docx, поддерживает билингвальные сценарии (диалог на одном языке + перевод в скобках под ним), и помогает аудитировать структуру по причинности и движению ценности. Скилл не привязан к конкретной истории — пользователь приносит свою.
development
Extract shot composition DNA from any car photograph into structured JSON — camera angle, lens, framing, lighting — stripped of car-specific details. Then reuse extracted angles with any car identity to generate new images at scale. Use when: extracting angles from reference photos, building a shot library, batch-analyzing car photography, replicating a great angle with a different car, running extraction pipelines in Freepik or Flora. Triggers: "extract this angle", "steal this composition", "shot DNA", "analyze this car photo", "replicate this shot with my car", "batch extract angles", "car photography analysis", "angle extraction", "build a shot library".