How to use the Japanese Seedance vocabulary agent skill

What the skill makes the agent do

When you ask for a Seedance 2.0 prompt in Japanese, the agent uses established Japanese cinematic terms rather than word-for-word translations. It writes concise clauses that follow a readable order: subject, action, camera, lighting, sound, and a preservation constraint. It can add a short English gloss when that helps you check the meaning.

The skill also keeps reference tags such as @Image1 exactly as you wrote them, and keeps user-supplied Japanese dialogue intact while it refines the visual direction. It is a language and wording aid. It does not generate video itself.

• Camera terms such as ゆっくりドリーイン (slow dolly in), 横移動のトラッキング (lateral tracking) and 低いアングル (low angle)
• Lighting terms such as 逆光 (backlight), 柔らかい窓光 (soft window light) and 輪郭光 (rim light)
• Motion and audio phrases such as 煙が薄く広がる (smoke spreads thinly) and 音楽なし (no music)
• A compact pattern for keeping a subject or product consistent while changing only motion, light or camera

Install the skill from the ZIP

Download the ZIP for this skill from this page and unzip it. The skill folder contains SKILL.md. To make it available across your projects, place the folder in ~/.claude/skills/. To limit it to one project, place it in that project's .claude/skills/ folder.

If you use the Claude apps, you can upload the skill instead. Open Settings, go to Capabilities, then Skills, and upload it there.

• Personal use across projects: ~/.claude/skills/
• Single project: the .claude/skills/ folder inside that project
• Claude apps: Settings → Capabilities → Skills

Check the requirements before you start

This skill depends on its parent skill, seedance-20, and on several referenced files: directors-read.md, surface-prompt-profiles.md and vocab/ja.md. These must be present in the repository for the full instructions to resolve. Without them, the skill cannot follow its references completely.

The skill does not need a Seedance 2.0 account to write prompts, but you need a Seedance 2.0 generation tool or account to actually render the prompts.

• The parent seedance-20 skill installed alongside this one
• The referenced files directors-read.md, surface-prompt-profiles.md and vocab/ja.md in place
• A Seedance 2.0 generation tool or account, only if you want to render the output

Trigger the skill with natural requests

The skill is meant to load when your request matches its description: Japanese Seedance 2.0 prompt wording, Japanese cinematic vocabulary, or translation of camera, lighting, action, VFX, audio and production terms into Japanese. You do not need special syntax. Describe what you want in plain language and mention Japanese.

Include your subject, the camera move you want, the light, any sound, and any reference tags you are using. The more concrete the shot, the more specific the Japanese wording.

• Write a Seedance 2.0 prompt in Japanese: a woman turns slowly toward a window, soft window light, quiet room tone
• Translate this camera direction into Japanese production terms: low angle, slow dolly in, backlight
• Keep @Image1 exactly as written and describe a product rotating with no logo distortion, in Japanese

Use the compact prompt pattern

The skill offers a compact pattern that keeps the prompt readable. Fill in the brackets with your own details. Keep the reference tag exactly as you wrote it, and keep only one camera move per prompt so the instruction stays clear.

Know the limits and review the wording

The author notes that the independent review artifact for this skill is empty, so the wording is treated as working production Japanese pending review by a locale specialist. Do not rely on it for native-level Japanese copywriting, or for legal or medical text.

Some of the skill's guidance is about tone and register. When you ask for dialogue, the skill keeps your lines and asks about relationship and tone only where that changes the brief. It does not change politeness levels to fit a timing target.

• Good fit: Seedance 2.0 prompts for Japanese-speaking audiences or bilingual projects
• Good fit: concise production Japanese for camera, lighting and audio cues
• Not a fit: native-level copywriting, legal text, medical text
• Not a fit: a complete Seedance workflow, since the parent skill covers broader guidance

Safety notes: scripts and network

According to the project's listing, this skill runs no scripts and uses no network access. It is made of instructions in a single SKILL.md file. You can read that file before installing to confirm what the agent will be told to do.

• Runs scripts: no
• Uses network: no
• Licence: MIT

Tips for better results

Keep reference tags exactly as written, including their spelling, case and punctuation. A written tag does not attach a file, so make sure your actual assets are bound to the roles you intend.

Ask for one clear camera move and one lighting setup per shot. If an adjective is vague, such as a style word, ask for the specific action, framing or light it should produce.

• Paste your exact dialogue lines so they are preserved unchanged
• Ask for an English gloss if you need to check the meaning
• Review any output that will be published in Japanese with a native speaker