film-storyboard-skill
Scannednpx machina-cli add skill RainLib/AI-Storyboard/film-storyboard-skill --openclawFilm Storyboard Skill
为 Storyboard Artist 提供专业的影视分镜方法论和提示词生成能力。
概述
此技能帮助将剧本转换为视觉化分镜，支持完整的三阶段流程：
- Beat Breakdown - 从剧本识别 9 个关键叙事时刻
- Beat Board (9 宫格) - 为每个 beat 生成视觉提示词
- Sequence Board (4 格) - 展开关键 beat 为连续镜头序列
核心原则：4C Framework
- Clear 清晰 - 每个提示词明确无歧义
- Concise 简洁 - 详细但不臃肿（Visual Description: 80-120 词，Lighting & Mood: 30-50 词）
- Consistent 一致 - 角色/场景/光色在所有 prompts 中保持一致
- Progressive 渐进 - 从 9 宫格到 4 格逐层细化
快速开始
生成 Beat Breakdown
从剧本中识别 9 个关键叙事时刻，包含：
- Beat 编号和标题
- 时间戳/页码
- 核心动作
- 情感价值
使用模板: templates/beat-breakdown-template.md
生成 Beat Board
为 Nano Banner（推荐）:
EPISODE {XX}: BEAT BOARD VISUAL SCRIPT
Beat 1: [标题]
Visual Description: [80-120 词视觉描述]
Lighting & Mood: [30-50 词光影氛围]
为 Midjourney:
每个 beat 独立 prompt + --ar 16:9 --style cinematic --v 6
生成 Sequence Board
选择关键 beat 展开为 4 个连续镜头，保持：
- 180 度轴线规则
- 角色外观一致性
- 光影连贯性
关键约束
严格禁止:
- ❌ Frontmatter 元数据在输出文件中
- ❌ 模板说明或注释
- ❌ "下一步"指令
必须包含:
- ✅ 角色规范描述（canonical description）
- ✅ 逐字重复的关键识别符
- ✅ 适当平台格式（Nano Banner/Midjourney/Gemini）
详细资源
方法论指南 📖
- storyboard-methodology-playbook.md - 完整分镜方法论
- 四大支柱详解
- Beat selection criteria
- 镜头构图和摄影
- 连贯性管理
- 高级电影技巧（蒙太奇、转场、时空处理）
提示词写法 📖
- gemini-image-prompt-guide.md - 提示词优化指南
- 叙事描述式风格
- 角色一致性技巧
- 光影描述方法
- Nano Banner 格式优化
平台和风格参考 📖
- REFERENCE.md - 平台特性和风格库
- Nano Banner vs Midjourney vs Gemini 格式对比
- 7 种视觉风格库（写实、动漫、概念艺术等）
- 6 种光影方案候选
- 宽高比选项
模板
- templates/beat-breakdown-template.md
- templates/beat-board-template.md
- templates/sequence-board-template.md
何时使用
自动触发场景:
- 用户请求"生成 beat breakdown"
- 用户请求"创建 beat board"或"9 宫格"
- 用户请求"生成 sequence board"或"4 格序列"
- Director 反馈需要修订视觉一致性
手动参考场景:
- 不确定提示词格式
- 选择视觉风格
- 解决角色一致性问题
- 优化平台特定格式
平台快速选择
- 追求一致性 → Nano Banner（一次生成完整 3x3 网格）
- 需要精细控制 → Midjourney v6
- 快速原型 → Gemini Imagen 3
用法: Storyboard Artist agent 自动引用此技能。方法论和指南（标记 📖）采用渐进式披露，仅在需要时查阅。
Source
git clone https://github.com/RainLib/AI-Storyboard/blob/main/.claude/skills/film-storyboard-skill/SKILL.mdView on GitHub Overview
This Film Storyboard Skill converts scripts into visual storyboards using a three-stage workflow: Beat Breakdown (9 key moments), Beat Board (9-panel visual prompts), and Sequence Board (4-frame sequences). It applies the 4C framework—Clear, Concise, Consistent, Progressive—so prompts stay unambiguous, visually cohesive, and adaptable across Nano Banner, Midjourney, and Gemini.
How This Skill Works
First, perform a Beat Breakdown to identify 9 narrative moments with titles, timestamps, core actions, and emotional values. Next, generate Beat Board prompts for each beat, delivering 80–120 word Visual Descriptions and 30–50 word Lighting & Mood sections. Finally, select key beats and unfold them into a 4-frame Sequence Board, maintaining a 180-degree axis, consistent character appearance, and coherent lighting across frames for cinematic continuity.
When to Use It
- Automated trigger to generate a Beat Breakdown from an input script
- Create a Beat Board (9-panel) with per-beat visual prompts
- Generate a Sequence Board (4-frame sequence) from key beats
- Director feedback requires revision to improve visual consistency
- Prepare platform-specific prompts for Nano Banner, Midjourney, or Gemini
Quick Start
- Step 1: From the script, identify 9 beats with titles, timestamps, core actions, and emotional values.
- Step 2: For each beat, generate a Beat Board prompt using 80–120 word Visual Description and 30–50 word Lighting & Mood, ensuring canonical descriptions.
- Step 3: Choose key beats and expand into a 4-frame Sequence Board, enforcing 180-degree axis, character consistency, and coherent lighting across frames.
Best Practices
- Apply the 4C Framework: Clear prompts, Concise descriptions, Consistent visuals, Progressive refinement from Beat Board to Sequence Board
- Use Visual Description (80–120 words) and Lighting & Mood (30–50 words) for each Beat Board prompt
- Provide canonical character and scene descriptions to maintain consistency across beats
- Include platform-specific formatting and syntax for Nano Banner, Midjourney, and Gemini
- Incorporate verbatim key identifiers across all prompts to preserve recognizability
Example Use Cases
- A drama script is parsed into 9 beats with titles and timestamps, producing a Beat Breakdown template and nine Beat Board prompts tailored for Nano Banner.
- Each beat’s 80–120 word Visual Description and 30–50 word Lighting & Mood are crafted to align with a cinematic tone and a consistent color palette.
- Beat prompts are exported for Midjourney with --ar 16:9 and cinematic style, one prompt per beat, ensuring uniform character design across frames.
- A selected subset of beats is expanded into a 4-frame Sequence Board, following the 180-degree axis rule and maintaining lighting continuity.
- Director feedback is applied by refining prompts to tighten visual consistency and adjust any mismatched lighting or costume details.
