cyberpunk-anime

Imported from okou-ai/vm0-skills (video-template/cyberpunk-anime/SKILL.md). Install upstream with npx skills add okou-ai/vm0-skills --skill cyberpunk-anime. Copyright stays with the author.

Cyberpunk Anime

A 2D cyberpunk anime style, not a fixed scene. Keep the user's character or subject exactly as briefed — and render it as hand-drawn anime in the neon-city look below. The style supplies the look; the user supplies the what. Tuned for Seedance (the platform's default video model): the prompt follows Seedance's subject → scene → motion → camera → light → style ordering, and the look is 2D animation (state this explicitly so the model doesn't drift to live action / 3D).

Video preview

1. Before generating a video, prepare a few keyframes matching the user's subject, style, and aspect ratio. Reuse suitable supplied or already approved images.
2. Show the actual images through accessible links, briefly describe the intended motion, and ask the user to confirm. End the turn and wait; do not start a video job before confirmation. Revise the preview if requested.
3. After confirmation, generate the video. Use the approved images as first/last frames or references when supported by the selected model; otherwise follow the approved visual direction in the prompt. Reuse existing confirmation for an unchanged preview.

Keep the preview message short: the images, a brief motion description, and one confirmation question.

What this style is

The essence: a lonely, beautiful neon-noir mood in hand-drawn anime — a small human moment dwarfed by a glowing, rain-soaked megacity. The goal is atmosphere and melancholy, the quiet feeling of a city at night, not action spectacle.

Touchstones: Blade Runner-flavored anime, lo-fi "chillhop" anime loops, Ghost in the Shell / cyberpunk OVA backgrounds, neon city night illustrations.

What makes it distinct: 2D cel animation + neon + rain + melancholy. Not live action, not 3D CGI, not warm or natural — flat cel shading, electric teal/magenta neon, wet reflective streets, a pensive lone figure.

Style dimensions (locked)

• Visual tone — neon cyberpunk: electric teal and magenta neon against deep night; glowing signage, wet reflective surfaces.
• Camera — dutch / slow drift: subtly tilted or slowly drifting framing; atmospheric, observational, not frenetic.
• Editing pace — fast cut (restrained): can cut between moody beats; keep it atmospheric, not action-frantic.
• Narrative mode — abstract / mood: vibe over plot; a feeling, not a story beat.
• Production type — 2D animation: hand-drawn cel-shaded anime. Explicitly not live action, not 3D CGI.
• Emotional tone — melancholic: pensive, lonely, bittersweet calm.
• Style reference — cyberpunk anime: neon megacity OVA backgrounds, rain and reflections.

Prompt construction

Write one cohesive video prompt in your own words, adapted to the subject. Hit these beats in Seedance order:

character/subject (as briefed) → neon rain-slicked city setting → quiet motion (rain, steam, drifting figure) → subtly tilted/drifting camera → teal-magenta neon glow → 2D cel anime style, melancholic

Always convey: 2D hand-drawn anime, cel-shaded · neon megacity backdrop with glowing signage · rain-slicked reflective streets · electric teal + magenta palette · a melancholic, lonely mood · atmospheric framing.

Never: live-action footage, 3D CGI, photorealism, warm natural daylight, bright cheerful tone.

Adapt the character and city details to the brief. Put aspect ratio, negatives, and seed in the params.

Generation parameters

• aspectRatio: 16:9 (cinematic anime) or 9:16 for vertical loops.
• resolution & model tier: pick a tier that supports the resolution — fast tiers cap lower (often 720p); 1080p needs a full tier.
• duration: 5–8s; let rain, neon flicker, and a slow drift carry it.
• negativePrompt: live action, photorealistic, 3D CGI render, warm daylight, bright cheerful tone, natural landscape, low resolution, extra fingers.
• generateAudio: on for rain/city ambience (pairs well with lo-fi music in edit).
• seed: mild lever for text-to-video; for look/character consistency use firstFrameImageUrl.
• firstFrameImageUrl: use the approved opening frame when the selected model supports this input.

How to apply

Render the user's subject as 2D cel anime in a neon rain-city. Always state the 2D-anime medium in the prompt (the model drifts to live action otherwise). If the brief wants photoreal or a bright/warm mood, this isn't the fit.

Worked examples

Same neon-anime look; the subject changes.

1. "a lone hacker" → a hooded figure looking down a rain-slicked neon alley, glowing signage, melancholic, slow drift (the picker thumbnail).
2. "someone eating ramen at night" → a lone figure at a glowing ramen stall, steam and neon reflections in wet pavement, pensive calm (see reference still).
3. "a girl on a train platform" → standing under flickering neon, rain outside, teal-magenta glow, subtly tilted framing.
4. "a cat on a fire escape" → a cat silhouetted against neon signs, rain, reflective puddles below.

Reference output

The reference still holds the neon-anime look on a different subject (ramen stall vs. hooded figure) — the style is subject-invariant. Use an approved frame of the user's subject when supported.