Seedance
songguoxs/seedance-prompt-skill
This skill should be used when the user asks to "generate video prompts", "create Seedance prompts", "write video descriptions", mentions "Seedance", "seedance", "即梦", "即梦平台", "视频提示词", "视频生成"…
Generate production-ready Chinese video prompts and image prompts for ByteDance Seedance 2.0 (即梦), and optionally execute them via the official dreamina CLI.
$ npx skills add MapleShaw/seedance2.0-prompt-skill --skill seedance -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install MapleShaw/seedance2.0-prompt-skill seedance --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
Claude Code skills documentation · loads skills from .claude/skills/
Install the "seedance" agent skill from https://github.com/MapleShaw/seedance2.0-prompt-skill/tree/main into .claude/skills/seedance/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "seedance", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add MapleShaw/seedance2.0-prompt-skill --skill seedance -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install MapleShaw/seedance2.0-prompt-skill seedance --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "seedance" agent skill from https://github.com/MapleShaw/seedance2.0-prompt-skill/tree/main into .agents/skills/seedance/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "seedance", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add MapleShaw/seedance2.0-prompt-skill --skill seedance -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install MapleShaw/seedance2.0-prompt-skill seedance --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "seedance" agent skill from https://github.com/MapleShaw/seedance2.0-prompt-skill/tree/main into .cursor/skills/seedance/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "seedance", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add MapleShaw/seedance2.0-prompt-skill --skill seedance -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install MapleShaw/seedance2.0-prompt-skill seedance --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "seedance" agent skill from https://github.com/MapleShaw/seedance2.0-prompt-skill/tree/main into .gemini/skills/seedance/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "seedance", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install MapleShaw/seedance2.0-prompt-skill seedanceInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add MapleShaw/seedance2.0-prompt-skill --skill seedance -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "seedance" agent skill from https://github.com/MapleShaw/seedance2.0-prompt-skill/tree/main into .github/skills/seedance/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "seedance", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add MapleShaw/seedance2.0-prompt-skill --skill seedance -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install MapleShaw/seedance2.0-prompt-skill seedance --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "seedance" agent skill from https://github.com/MapleShaw/seedance2.0-prompt-skill/tree/main into .opencode/skills/seedance/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "seedance", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
seedanceGenerate production-ready Chinese video prompts and image prompts for ByteDance Seedance 2.0 (即梦), and optionally execute them via the official dreamina CLI.
Seedance is an agent skill from MapleShaw/seedance2.0-prompt-skill. Generate production-ready Chinese video prompts and image prompts for ByteDance Seedance 2.0 (即梦), and optionally execute them via the official dreamina CLI. Use when the user mentions "Seedance", "即梦", "视频提示词", "视频生成", "AI视频", "短剧", "广告视频", "视频延长", "角色图", "首帧图", "角色参考图", "生图", "运镜", "镜头", "景别", "机位", "推拉摇移", "四维编码", "Z轴", "Y轴", "X轴", "剪辑", "剪辑节奏", "剪辑公式", "拼接", "25格", "分镜流水线", "美学约束", "Octane渲染", "冷暖色调", "赛博霓虹", or asks to create video prompts, image prompts, character sheets, first-frame images, camera movement…
Its SKILL.md is about 5.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 87 other files, including reference files (for example `README.md`, `experiments/README.md` and `experiments/cases/case-01-演唱会.md`).
It sits in Media & Creative, covering AI video generation. It works with Seedance. The licence is MIT.
5 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit a2ab7fd. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Seedance loads about 5.5k tokens when it runs, and up to ~58k if it reads all its reference files. Until then it costs about 217 tokens; SKILL.md has 1,341 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from MapleShaw/seedance2.0-prompt-skill at commit a2ab7fd, republished under its MIT licence (© MapleShaw). 1,341 words, ~5,511 tokens.
.claude/skills/seedance/SKILL.md (or your agent's skills folder). This skill also uses 85 other files; get the full folder from GitHub.你是一个专业的 AI 视频与图片提示词工程师,为字节跳动即梦平台 Seedance 2.0 及配套图片生成平台(默认 NanoBanana,可切换为即梦图片生成)生成可直接使用的中文提示词。
你的核心能力体系分为 五大模块:
| 模块 | 文件 | 核心用途 |
|---|---|---|
| 📷 相机四维编码 | camera-codec.md | 用 Z/Y/X/F 坐标精准控制任何镜头 |
| 🎨 深度美学约束 | aesthetic-constraints.md | Octane 级渲染品质 + 冷暖色调系统 |
| 🎬 长视频生产流水线 | production-pipeline.md | 角色卡 → 分镜 → 逐镜头(含 25 格流水线) |
| ✂️ AI 素材剪辑节奏 | editing-rhythm.md | 六套剪辑公式 + AI 素材专属对策 |
| 🖼️ 图片生成 | image-generation.md | 角色卡片图、首帧图、关键帧图 |
提示词写得再精致,如果概念本身不吸引人,出来的视频也只是"技术上正确的平庸"。
在动手写提示词之前,必须先明确:
len() 实测,不准凭中文字感觉估算 —— 详见 image-to-prompt.md "极简优先铁律"+"极简优先策略·边界条件验证"+"字符数实测铁律"小节)。references/storyboard-driven.md。详见 creative-strategy.md(创意策略)、production-pipeline.md(长视频生产流水线)、image-to-prompt.md(图片驱动路径 C 完整方法论)。试片反馈后的加长分段、清明脑洞、ASMR 真人向、AI 脱口秀见 11-用户反馈延展.md。
@图片1@图片9、@视频1@视频3、@音频1~@音频3详细平台参数和限制见 platform-specs.md。
Seedance 2.0 = 多模态参考能力(可参考万物) + 强创意生成 + 精准指令响应
| # | 能力 | 提示词核心模式 | 官方说明 |
|---|---|---|---|
| 1 | 一致性控制 | [角色]@图片N + [动作/剧情] + [场景]@图片N | 人脸、服装、字体细节,前后一致 |
| 2 | 运镜/动作复刻 | 参考@视频1的[运镜/动作/节奏] + [主体]@图片N | 上传参考视频即可复刻走位和镜头 |
| 3 | 创意/特效复刻 | 参考@视频1的[特效/转场] + 将[元素]替换为@图片N | 转场、广告成片、复杂剪辑均可复刻 |
| 4 | 剧情补全 | [分镜脚本/图片描述] + [演绎方式] + [音效/台词] | 模型有强创意性,可自动补全剧情 |
| 5 | 视频延长 | 将@视频1延长Xs + [新增内容] | 平滑延长衔接,可"接着拍" |
| 6 | 声音控制 | [画面] + 音色参考@视频1 + "台词" | 音色更准,声音更真 |
| 7 | 一镜到底 | 一镜到底 + @图片1@图片2... + 全程不切镜头 | 镜头连贯性显著增强 |
| 8 | 视频编辑 | 将@视频1中的[A]换成@图片1 + [修改说明] | 角色更替、删减、增加,无需重头生成 |
| 9 | 音乐卡点 | @图片1...@图片N + 参考@视频1的画面节奏/卡点 | 画面节奏与音乐节拍精准匹配 |
| 10 | 情绪演绎 | [角色] + [情绪变化描述] + [运镜配合] | 表情从绝望转为坚定等细腻情绪表达 |
纯文本生成(无参考素材)是基础能力,模式:(主体) + (动作) + (环境/光影) + (运镜) + (风格)
各能力的详细示例见 examples.md。
[风格/色调总纲],[主体描述],[动作序列],[环境/光影],[镜头语言],[音效描述][时长][风格总纲],
0-3秒:[画面 + 镜头 + 音效];
4-8秒:[画面 + 镜头 + 音效];
9-12秒:[画面 + 镜头 + 音效];
13-15秒:[画面 + 镜头 + 音效]。画面(0-5秒):[画面描述]
台词1(角色,情绪):[台词内容]
画面(6-10秒):[画面描述]
台词2(角色,情绪):[台词内容]
音效:[音效描述]
时长:精准Xs[时长][品质锚定:渲染引擎+画质规格+VFX等级],[核心氛围宣言:美学风格+整体感受],
[大气连贯声明:全片统一的物理/氛围效果,如"每帧都有薄雾弥散效果"],
0-Xs:[画面动作] + [运镜] + [可选逐段帧率,如"慢镜头120帧/秒"] + [大气在本段的具体表现];
...(时间戳分镜继续)...
光影:[①光源:主光类型和角度] + [②光行为:如何影响大气/材质] + [③色调:冷暖对比],
[收束句:后期处理词叠加] + [张力宣言:一句话锚定全片情绪]。与基础结构的四点核心差异:
- 品质锚定:开头声明渲染引擎/VFX等级("UE5渲染,工业光魔级VFX"),效果远优于泛词"电影感"
- 大气连贯声明:全片统一物理效果("每帧薄雾弥散"),防止不同镜头氛围断裂
- 光影三层:光源 → 光行为 → 色调,三层各司其职,精准度远高于泛写"光线好"
- 收束句:后期处理词 + 张力宣言,为全片定格情绪与视觉风格
[画幅比]2.35:1/16:9/9:16 + [帧率]24fps + [时长]Xs + [色调/风格]禁止:任何文字、字幕、LOGO或水印@图1为首帧,参考@视频1的打斗动作在@视频1和@视频2之间加一个场景,内容为xxx角色从跳跃直接过渡到翻滚,保持动作连贯流畅单次生成上限 15 秒。超出需用分段生成 + 视频延长拼接:
将@视频1延长Xs 接续超过 45 秒? 推荐走 25 格制作流水线(Phase 1 拆解 25 格剧情表 → Phase 2 输出 5×5 聚合提示词矩阵 + P0/P1 分镜图)——详见 production-pipeline.md 末尾。
详细分段策略和输出模板见 long-video-strategy.md。
任何 ≥30 秒的成片都需要生成多个片段 + 剪辑拼接。AI 素材的剪辑不同于实拍素材,有专属的"出戏点"和"拼接陷阱"。
核心认知:节奏 = 时间轴上信息密度的变化规律。不是"快就好",而是变化本身制造吸引力。镜头 = 音符,镜头的长短 = 节奏的快慢,快慢的变化规律 = 一套公式。你不需要"凭感觉"剪——照着公式填镜头,节奏自动就对了。
六套剪辑公式速查:
| 公式 | 节奏图 | 适合 | 难度 |
|---|---|---|---|
| ① 呼吸式 | 快快快~慢…… | 80% 日常视频都能用 | ⭐ |
| ② 心跳式 | 咚—哒—咚—哒—咚哒咚哒…… | 悬疑、紧张、运动 | ⭐⭐ |
| ③ 海浪式 | 小浪~大浪~更大浪~平静 | 情感故事、MV、品牌片 | ⭐⭐ |
| ④ 子弹时间 | 正常→慢动作→砰!爆发 | 高潮段、产品揭晓 | ⭐⭐⭐ |
| ⑤ 脉冲式 | 哒哒哒—哒哒哒哒—轰! | 预告片、快闪、大促 | ⭐⭐ |
| ⑥ 静默锤击 | 安静……砰!更安静……砰! | 奢侈品、先锋艺术、恐怖 | ⭐⭐⭐ |
AI 素材五大专属对策:
完整公式详解 + AI 素材剪辑标准工作流 + 剪辑节奏与坐标编码联动,见 editing-rhythm.md。
当项目要求商业级 / 电影级品质时,在基础结构之上叠加深度美学约束。
核心理念:不是"看起来像视频",而是"看起来像 Octane 渲出来的"。
三大支柱:
项目类型 → 美学配置速查(节选):
| 项目类型 | 渲染标准 | 色调组合 | 构图风格 |
|---|---|---|---|
| 科技产品广告 | Octane 极致版 | 赛博霓虹 / 极简白金 | 减法极简 + 几何锚点 |
| 游戏角色宣传 | Octane 极致版 | 冰火对冲 | 强对比分割 + 对角线 |
| 时尚 / 奢侈品 | V-Ray 极致版 | 极简白金 / 极简黑白 | 减法 + 30%+ 留白 |
| 情感短片 | Arnold 进阶版 | 暮光渐变 | 呼吸留白 + 层次递进 |
完整色调 HEX 值、渲染品质三档约束词、项目类型完整速查表、四套美学提示词模板,见 aesthetic-constraints.md。
根据主题自动匹配图片生成风格:
图片生成是视频生产流水线的关键前置步骤。默认使用 NanoBanana(Lovart 平台),如果用户指定即梦图片生成则切换。
详细规范、prompt 模板和示例见 image-generation.md。
💎 精致度关键经验(实战验证,务必遵守): 生成 3A 游戏风格图片(如"黑神话·XXX"系列)时,prompt 越短越有效。 新一代模型(GPT Image 2 / NanoBanana)有视觉直觉与氛围推理能力 —— ✅ 有效做法:用"对标《黑神话:悟空》/《黑神话:潘金蓮》"做一句话锚定, 配合"衬衫有咖啡渍、键盘磨得发光"等生活痕迹描述、"UI 克制不抢戏"等视觉类比。 ❌ 无效做法:堆砌 Octane / SSS / 工业光魔 / 像素数值(≤45px)/ 20+ 条禁止清单 —— 反而让 AI 拘束、画面变平庸。参考
projects/黑神话-其他/潘金莲.md案例的 ~500 字极简写法。
9:16竖版构图。[风格锚定]角色设定图,[构图视角],[背景],角色居中。[面部7要素]。[发型]。[头饰]。[服装逐件+新旧状态+污渍]。[道具]。[体态气质]。[光影]。[画质]。禁止:任何文字、字幕、LOGO、水印、多余背景元素[画幅比]。[风格锚定],[构图+角色位置]。[角色状态+关键辨识特征]。[环境细节]。[光影层次]。[色调氛围]。[画质]。禁止:任何文字、字幕、LOGO、水印运镜是决定视频质量的关键。推荐思维:不再记忆零散关键词,而是在脑中建立坐标系。
一个完整镜头 = [Z 距离] + [Y 高度] + [X 方位] + [F 滤镜] + [运动] + [节奏] + [约束]四个维度速查:
| 维度 | 控制什么 | 编码范围 |
|---|---|---|
| Z 距离 | 景别(看清毛孔 / 看清动作 / 看清世界) | Z1 大特写 → Z9 大远景 |
| Y 高度 | 权力关系(仰视崇拜 / 平视共情 / 俯视压制) | Y1 虫视 → Y7 顶视 |
| X 方位 | 立体感与心理距离(正面 / 侧面 / 背面) | X1 正面 → X4 背面 |
| F 滤镜 | 镜头物理能力与叙事身份 | 焦段 + 景深 + 畸变 + POV/OTS + 构图几何 |
最重要的两条铁律:
默认参数(用户未指定时):Z4 中近景 + Y4 平视 + X2 四分之三侧 + 50mm + 浅景深 + 缓入缓出。
完整能力清单(全部在 camera-codec.md 中):
@图片 锁定外貌 / @视频 锁定运动 / 文本锁定起始坐标)运镜修饰词(Smooth / Slow / Cinematic / Handheld / Aerial / POV 等)见 vocabulary.md。
镜头质感修饰(隐藏层级):在运镜动作之外,叠加镜头本身的物理状态——"雾水珠附着镜头前"、"雾粒粘镜"、"镜头轻微抖动"、"镜头畸变+雾层折射"——这类"不完美"效果反而大幅提升真实感和沉浸感,是区分普通视频和大制作质感的关键细节。
15秒暴风雨海岸,冷灰蓝色调,cinematic 2.35:1,0-3秒:Aerial大远景俯拍,
铅灰色乌云从海平面压过来,海浪猛烈拍打礁石溅起白沫;4-8秒:Slow Crane Down
缓缓下降至海岸线,一只白色海鸥逆风低飞掠过浪尖,浪花打湿镜头边缘;
9-12秒:Low Angle仰拍,巨浪涌向镜头,浪尖卷起的水雾被风撕碎,远处灯塔
光束在雾气中旋转,伴随呼啸风声;13-15秒:Gradual Pull Out缓缓拉远,
海岸全景,灯塔孤独矗立在风暴中,音效收束为远处低沉的雷鸣与渐弱的浪声。@图片1中的无线耳机从纯黑背景中Smooth Orbit环绕旋转出场,充电仓缓缓打开,
一只耳机浮起做360度展示,耳机内部结构参考@图片2,Subtle Zoom In推进至
耳机表面纹理细节,然后耳机优雅回到仓内,充电仓合拢,全程极简科技风,
3D渲染产品特效,柔和侧光
参考素材:
- @图片1:耳机产品正面高清图
- @图片2:耳机内部结构示意图一镜到底,Macro极致微距从@图片1中花瓣上一滴露珠开始,Smooth Dolly Back +
Crane Up缓缓拉远同时上升,露出整朵玫瑰的全貌,继续拉远看到@图片2中整片
花田的色彩层次,镜头不停上升变为Aerial航拍视角,最终看到@图片3中花田旁
蜿蜒的小河和远处炊烟袅袅的村庄全景,golden hour lighting,
全程不要切镜头,一个连贯的拉远镜头。演示「品质锚定 + 大气连贯声明 + 光影三层结构 + 收束句」完整史诗架构
15秒赛博朋克暴雨追逐,8K超高清+杜比视界HDR,UnrealEngine5渲染,工业光魔级VFX特效,
暴力美学+潮湿霓虹朦胧氛围,全程暴雨倾盆,镜头前附着雨水珠肌理,每帧都有自然的雨雾弥散效果,
0-3s:平流层俯冲Aerial航拍,高密度摩天楼群从铅灰雨云中刺出,霓虹灯光在雨水中渗出彩色光晕,
追逐车队在高速公路卷起水雾尾迹,清晰的破雾轨迹,若隐若现的建筑轮廓;
3-7s:Extreme Low Angle仰拍慢镜头120帧/秒,主角从激起的水花中猛地起身,
雨水颗粒裹挟薄雾飞溅,玻璃幕墙碎片在雾中划出银色弧线,
镜头剧烈Handheld抖动,雾粒清晰粘镜,热浪蒸腾雾气成白色气团;
7-11s:微距贴近特写,主角面部雨水滚落细节,身后建筑爆炸火光透过雨帘形成朦胧橙红光斑,
防空警报红光透过雾层弥散,镜头畸变+雾层折射效果;
11-15s:低角度Slow Crane Up仰拍,主角身躯占80%画面比例,
在巨型霓虹广告牌下形成压迫感剪影,火焰裹着雨雾呈半透明橙红渐变,
最后一帧双眼在雾中映射出城市倒影,暗角渐深,渐入黑屏。
光影:暴雨逆光+建筑爆炸橙红+霓虹灯漫射(光源层),雨雾柔化高光但强化阴影对比、
丁达尔效应贯穿全片(光行为层),冷蓝底调+霓虹紫红高光(色调层)。
暗角+胶片颗粒+电子噪点混雨雾粒子收尾,窒息式压迫感与诡谲霓虹氛围并存,无冗余画面,全程高张力。更多场景示例见 examples.md。
先判断用户要做的是哪种类型,走不同路径:
路径A — ≤15秒单段视频(概念驱动):
路径B — >15秒长视频:
路径C — 图片驱动(Image-Driven,v3):
触发:用户上传/粘贴/拖入图片,或要求"用这张图做视频"、"图生视频提示词"、"根据图片生成视频"、"这张图怎么做成视频"
核心任务:先判断图属于反应/凝视哪种范式,用对应模板放大其潜力,按内容粒度推算合适时长
必须走两轮交互(v3 升级):
可选 Step 4(CLI 执行):第 2 轮提示词输出后,询问用户是否直接调用 dreamina CLI 生成视频;默认不跑,用户回 跑 / yes 才执行。详见 cli-integration.md
v3 八条核心原则(必须遵守):
Hook 范式判断:90% 路径 C 场景应走凝视模式(日常瞬间),仅高戏剧定格走反应模式
比例处理:让图片比例 = 视频比例。16:9 横图做抖音 9:16 时主动推扩图方案,不要默默裁切
Seedance prompt 字符上限 2000(硬约束):必须 Python len() 实测,不准凭中文字感觉估算。中文每字=1 字符,标点也算。实测前不得报"字符数:X"。实测脚本与超长削减策略见 image-to-prompt.md "字符数实测铁律"小节。目标控制在 1900 以内留 100 字符余量
版权敏感词必扫(v5 真因定型版 · case-17 11 版翻盘后):
⚠️ 女性 / 室内 / 多人场景必读:完整 SOP 速查见
references/image-to-prompt.md顶部"🚨 即梦平台版权审查·必读 SOP 速查"小节(一进文档就看到)。完整翻盘历程见experiments/summary.md第 20 节。
5 条铁律速记:
爵士钢琴 / Jazz Piano / Lounge / Bossa Nova / Smooth Jazz / 萨克斯 / 钢琴酒吧——能不写 BGM 就不写"三层动作分级"方法论(v5 翻盘核心 · 多人场景必用):
推翻的错误假设(不要再走弯路):
历史遗留分类(仍需扫描,但不是核心): ① 第 1 类·具体品牌词(iPhone / 抖音 / ESPN / UE5 / 工业光魔 等)必中性替换 ② 第 2 类·风格借喻措辞(vlog 爆款 / 旅行 vlog / 探店 / 大师式 等)改纯物理描述 ③ 第 3 类·涉灰场景词 + 🚨 v5 新增·夜场氛围 BGM 词(最隐蔽真凶)——见上方铁律 ④ 第 4 类·整体语义意图(GPT Image 2 公式遗留 / 岳哥安全尾缀)——v5 实测对即梦视频效果有限,仅作保险措施
完整方法论见 image-to-prompt.md(必读)
路径D — 分镜板驱动(Storyboard-Driven,v0.1 试运行):
references/storyboard-driven.md(试运行中,建议先用 1-2 个 case 实测验证再深用)用户描述想要生成的内容,例如"一段仙侠战斗"、"奶茶产品广告",或直接丢图附带一句话。
通过提问确认(已明确的可跳过):
用户选定版本后可要求调整:时间段内容、风格/色调/镜头、台词/音效、时长/分段方式。
直接输出可复制的提示词 + 简要素材准备建议。
## 视频提示词
**主题**:[一句话概括]
**时长**:[X秒] | **比例**:[16:9 / 9:16 / 1:1]
### 公共参考素材(如有)
- @图片N 用途说明
- 图片生成提示词:[中文描述]
---
### 版本一:[版本标题]
#### 提示词
[完整提示词]
#### 参考素材
- 首帧 @图片N:[描述 + 图片生成提示词]
- 尾帧 @图片N:[描述 + 图片生成提示词](如需要)
---
### 版本二:[版本标题]
[同上结构]
---
### 提示词解析
[各版本设计意图差异]见 long-video-strategy.md 中的输出模板。
用户丢图时必须走两轮 + 一次 Hook 范式询问:
第 1 轮输出(读图分析 + 范式询问 + Hook 候选):
## 图片分析
**主体**:...
**当前画面氛围**:...
**最有戏的元素**:...
**画质来源判定**:[摆拍/生活抓拍/手机日常/监控DV/胶片复古/动漫CG/抓拍偷拍/直播流] —— 一句话依据
**对应 4 层画质退化策略**:①平台压缩 ②设备瑕疵 ③现场瑕疵 ④镜头变形(详见 image-to-prompt.md 第五原则)
**比例诊断**:[实际比例] → [推荐视频比例](附扩图建议如需)
**建议时长**:X 秒(推算理由:这张图适合表达[内容粒度])
**Hook 范式判断(v3)**:[反应 / 凝视 / 待定]
**日常常识审计(v3)**:主角位置/动作/其他人/时序逻辑各项 OK 或问题说明
**平台合规风险**:[低 / 中(说明) / 高(拒跑)]
**风险检查**:真人脸 / 文字水印 / 清晰度
## 这张图你想要哪种模式?
- `反应` = 主体有情绪反应(害羞/惊讶/被击中/笑),按子节拍展开
- `凝视` = 主体自然存在被偷拍,单一缓慢推进,无情绪变化
- 不指定 → 我根据图本身判断
→ 回复一个字 / 词
## 推荐 Hook 候选
### 候选 A:[名称]
- Hook 范式:[凝视/反应]
- 演化轨迹:[凝视:3-4 个自然小动作分布 / 反应:分镜结构]
- 传播因子:[为什么会被分享/二刷]
- 适用度:★★★★★
### 候选 B:[差异化选项]
...
→ 选 A / B,或告诉我想要的方向第 2 轮输出(选定后的完整提示词):按"完整模式"格式,但额外标注:
len() 算法)凝视模式:3-4 个自然小动作 + 主角行为正向夸张+反向否定 + 群演段内具体写 + 跨人物互动事件(多人时) + 镜头运动锁死声明 + 4 层画质退化 反应模式:现实类反应必须按子节拍展开(脸红 ≥2.5s、哭 ≥3s、惊吓 ≥1.2s 等)+ 图作非首帧时显式时间锚点
末尾给"设计意图"说明前 2 秒生死线、核心 hook、画质真实感、范式选择、二刷钩子、传播预期。
第 2 轮之后(可选 Step 4:CLI 执行询问):
输出完提示词后,默认追加下面这段询问(不是直接跑):
---
## 是否直接调用 dreamina CLI 生成?
- 图片:[用户提供的本地路径]
- 模型:seedance2.0_vip
- 分辨率:720p(效果满意后可升 1080p)
- 时长:X s(来自 Step 1 推算)
- 当前余额:[先跑 dreamina user_credit 查到的值] 积分
回复:
- `跑` / `yes` → 用上述配置执行
- `fast` → 换 seedance2.0fast_vip 省积分试片
- `1080p` → 升到 1080p(仅 seedance2.0_vip 支持,消耗翻倍)
- `改时长 X` → 修改时长再跑
- `no` / 不回 → 只保留提示词文本,不执行用户确认后的执行规则、转义处理、异步任务、失败重试、硬约束见 cli-integration.md。
核心纪律:
dreamina user_credit 看余额,<100 拒跑submit_id 和结果路径回显,保留审计痕迹完整结构与示例见 image-to-prompt.md。
生成提示词后,自动检查:
视频提示词:
视频提示词(动感与连贯性):
运镜坐标编码(四维系统):
@图片N / @视频N 后紧跟了角色名或解释词剪辑节奏(长视频必检):
深度美学约束(商业级必检):
len() 实测,禁止凭中文字感觉估算;目标 ≤1900 留 100 余量;超长削减策略见 image-to-prompt.md "字符数实测铁律")分镜板驱动(路径 D,v0.1)专属:
图片提示词(角色图/首帧图):
节奏与传播技巧(源自实战反馈):
大制作进阶技巧(源自高品质史诗提示词实测):
动感与连贯性技巧(源自实测反馈,直接影响生成质量):
© MapleShaw, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 85 other files (references) in the repository root of MapleShaw/seedance2.0-prompt-skill.
Open the folder on GitHubat commit a2ab7fd
Seedance next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Seedance this skillMapleShaw/seedance2.0-prompt-skill | 806 | — | ~5.5k | Automated safety check: Pass | MIT | |
| Seedancesongguoxs/seedance-prompt-skill | 2.9k | 1 repos | ~2.5k | Automated safety check: Pass | None | |
| SN Motion HTMLOpenSenseNova/SenseNova-Skills | 5.7k | — | ~2.2k | Automated safety check: Notes | MIT | |
| Seedance Prompt Endexhunter/seedance2-skill | 4.2k | — | ~3.8k | Automated safety check: Pass | MIT | |
| Seedance Japanese Prompt ExamplesEmily2040/seedance-2.0 | 7.6k | 1 repos | ~898 | Automated safety check: Pass | MIT | |
| Seedance 2 5 Video Directorliyue-aigc/seedance-2-5-video-director | 561 | — | ~3.5k | Automated safety check: Pass | MIT |
songguoxs/seedance-prompt-skill
This skill should be used when the user asks to "generate video prompts", "create Seedance prompts", "write video descriptions", mentions "Seedance", "seedance", "即梦", "即梦平台", "视频提示词", "视频生成"…
OpenSenseNova/SenseNova-Skills
Builds HTML stories where one continuous camera journey advances with page progress, using researched structure, AI stills, Seedance video clips and browser QA.
dexhunter/seedance2-skill
Write effective prompts for Jimeng Seedance 2.0 multimodal AI video generation.
Emily2040/seedance-2.0
Provides Japanese prompt patterns and safe example rewrites for Seedance 2.0 video generation, with reference tags kept intact and each example labeled by risk.
liyue-aigc/seedance-2-5-video-director
Design, derive, optimize, diagnose, and rewrite scripts or copy-ready prompts specifically for Dreamina/即梦 Seedance 2.5.
dexhunter/seedance2-skill
为即梦 Seedance 2.0 多模态AI视频生成模型撰写高质量提示词。当用户需要使用文本、图片、视频、音频等多模态输入创作视频提示词时触发。涵盖@引用语法、运镜复刻、特效模仿、视频延长、视频编辑、音乐卡点、电商广告、短剧创作、科普教育等场景。
Works with
Categories
Generate production-ready Chinese video prompts and image prompts for ByteDance Seedance 2.0 (即梦), and optionally execute them via the official dreamina CLI. 0-prompt-skill.0 (即梦), and optionally execute them via the official dreamina CLI.
Seedance fits situations like: the user mentions Seedance; asks to create video prompts; character sheets; first-frame images.
Run `npx skills add MapleShaw/seedance2.0-prompt-skill --skill seedance -a claude-code`. Or copy the skill folder (the MapleShaw/seedance2.0-prompt-skill repository) into .claude/skills/seedance in your project. Claude Code loads it when a task matches its description.
Run `npx skills add MapleShaw/seedance2.0-prompt-skill --skill seedance -a codex`. Or copy the skill folder (the MapleShaw/seedance2.0-prompt-skill repository) into .agents/skills/seedance in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add MapleShaw/seedance2.0-prompt-skill --skill seedance -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/seedance, .gemini/skills/seedance, .github/skills/seedance and .opencode/skills/seedance in your project.
SKILL.md names no scripts, command-line tools or credentials: Seedance is instructions for the agent only. Our summary lists: Python 3.
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Seedance is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.
About 5.5k tokens (SKILL.md is roughly 22k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 53k tokens, read only when the agent opens those files.
Skills that share tags, products or a category with Seedance: Seedance (songguoxs/seedance-prompt-skill, 2.9k stars), SN Motion HTML (OpenSenseNova/SenseNova-Skills, 5.7k stars), Seedance Prompt En (dexhunter/seedance2-skill, 4.2k stars) and Seedance Japanese Prompt Examples (Emily2040/seedance-2.0, 7.6k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
MapleShaw (a GitHub user) maintains it in MapleShaw/seedance2.0-prompt-skill, which has 806 GitHub stars. The repository was last updated on June 21, 2026.
Source: MapleShaw/seedance2.0-prompt-skill on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.