How to use the Chinese Seedance Prompt Vocabulary agent skill

What is an agent skill and what does this one do?

An agent skill is a folder containing a SKILL.md file that an AI agent loads when a task matches the skill's description. This skill's description says it should be used when you ask for Chinese Seedance 2.0 prompt wording, Mandarin cinematic vocabulary, Chinese prompt compression, or translation of camera, lighting, action, VFX, audio, and production terms into Chinese.

When it is active, the agent writes concise Chinese prompt text, adds an optional English gloss when useful, and keeps reference tags unchanged. It is a text-writing guide only. It does not generate video itself.

• Camera wording such as 缓慢推镜, 横向跟拍 and 固定中景
• Lighting wording such as 侧逆光, 柔和窗光 and 轮廓光
• Motion and audio phrases such as 慢慢转身, 安静环境声 and 无配乐
• First/last-frame phrasing such as @图片1 为首帧 and 自然过渡到尾帧

Install the skill from the ZIP download

Download the ZIP for skills/seedance-vocab-zh from this page. Unzip it into one of two locations: your personal skills folder at ~/.claude/skills/, or a project-specific .claude/skills/ folder inside the project you are working in.

If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills. The skill's package contains a single file, SKILL.md.

• Personal install: ~/.claude/skills/
• Project install: .claude/skills/ inside your project
• Claude apps: Settings → Capabilities → Skills

Trigger the skill with a natural request

You do not need a special command. Ask for what you want in plain language and mention Chinese, Mandarin, or Seedance 2.0. The skill's description lists the kinds of requests it covers, so requests such as writing a Chinese prompt, compressing a Chinese prompt, or translating camera and lighting terms into Chinese should match it.

The source also states that the skill can be invoked by the user (user-invocable: true), but it does not list a specific slash command, so describe the task in your own words.

• Write a Seedance 2.0 prompt in Chinese for a rainy street scene with a slow push-in camera.
• Translate this English Seedance prompt into Mandarin and keep @图片1 as the first frame.
• Compress this Chinese prompt to be shorter while keeping the camera and sound direction.
• Give me Mandarin terms for a cold moonlight lighting setup.

Use reference tags correctly

Reference tags such as @图片1 must stay exactly as written. The skill says to preserve the actual token, including its script, case, spacing and punctuation, and that a written example token does not attach a file. Bind each placeholder to a real asset by its requested role, such as first frame or last frame.

The compact pattern the skill uses looks like this. Replace the bracketed parts with your own subject, action, light and sound.

Keep the Simplified or Traditional choice deliberate

The skill treats the prompt script, subtitle script, locale and intended voice as separate decisions. It does not force Simplified Chinese, and it does not infer a voice from the written script. If you have not chosen a script, the agent may ask one focused question.

The skill forbids converting text by find-and-replace. For example, 头发 should become 頭髮 in Traditional, not 頭發. Keep exact dialogue and reference tags unchanged, and convert only the text you asked to adapt.

Review the output before you generate

The skill's own notes say the shipped independent review artifact is empty, and that the wording is working production wording pending locale-specialist review. Treat the output as a strong draft. Read it yourself, check that each phrase points to a visible action, camera behaviour, light source, sound or constraint, and check every reference tag against your uploaded assets.

For final locale, voice-over or dialogue decisions, use a native Chinese specialist.

• Confirm each reference tag matches your asset and its role
• Remove any adjective that does not change the shot
• Check that dialogue text is exactly what you intended