Vigen — AI skill for Claude Code
Claude Code skill that turns a one-line idea into a complete AI video production plan: storyboard, shot design and copy-ready prompts for gpt-image / happyhorse-i2v / cosyvoice, with character, scene.
How to install Vigen
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open ottolsq/vigen and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Vigen does
Claude Code skill that turns a one-line idea into a complete AI video production plan: storyboard, shot design and copy-ready prompts for gpt-image / happyhorse-i2v / cosyvoice, with character, scene & style consistency anchoring. 一句创意 → 剧本·锚定表·图片/视频提示词·台词·剪辑清单。
Alternatives in AI
- Guizang Ppt Skill — AI-agent Skill for generating polished HTML slide decks: editorial magazine and Swiss layouts, image prompts 24.9k ★
- Blender AI MCP Addon — AI assistant and MCP server for Blender: visual chat, scene tools, multi-agent workflows 531 ★
- Agency Orchestrator — 🚀 One sentence → multi-AI-role collaboration → complete plan in minutes 453 ★
README
vigen — AI 视频制作辅助 Skill(ai-video-maker)
把一句创意或一份剧本变成一套完整的、可直接复制到生成工具使用的视频制作方案。面向零影视基础的用户:分镜、运镜、转场、光影、提示词工程全部自动补全,只需回答需求问卷并在锚定确认点(GATE1)做决定。
它解决什么问题
AI 视频制作的痛点不在"没有工具",而在"不知道怎么指挥工具":
| 痛点 | 本 Skill 的解法 |
|---|---|
| 图片有"AI 油腻感"、单一死板 | "描述这张照片怎么拍出来的"方法论 + 画面张力四法(明暗纵深/尺度对比/框中框/曝光分区)(realism.md) |
| 分段生成色调漂移 | 全片色彩基调:3-5 个十六进制色号按面积分配(环境底色60%/中间调25%/点缀色10%只给特定物体/肤色基准/暗部色),色号+落点写进每条提示词(styles.md) |
| 提示词精准度不够、模型听不懂 | 语义绑定三层法:绑位置关系不绑坐标 / 绑因果不绑并列 / 绑先验概念包不对抗先验("血月"胜"红色的月亮");正向引导>负面约束(prompt-rules.md 四大底层认知) |
| 复杂动作怎么都生成不出来 | 先验壁垒认知:训练数据没见过的姿势纯文字写不出来——先用首帧图锁定姿势再让它动;分镜表标注反先验动作 |
| 人物/场景/道具跨图不一致 | 四层锚定:角色/场景/道具/风格圣经逐字复用 + 参考图多图输入 + 参考图六条硬标准(短边≥1024/纯色背景/同一套光/中性表情/一张脸/脸身分离,一张图只回答一个问题)(consistency.md) |
| 道具状态混乱、物品凭空出现 | 道具六条判定规则(≥2镜出现/触碰变状态/会位移/剧情承载/镜头要对焦到的物品或图案/角色强关联随身物)→ 独立参考图每状态一张 + CONTINUITY RULE 连续性声明 |
| 首帧图生视频动作突兀 | 情景图强制捕捉"蓄力瞬间",视频从静止自然起势 |
| 动作违反物理常识 | 动作三段式(起始→过程+速度曲线→明确终点)+ ENDING FRAME 明确尾点 |
| 跨片段音色漂移 | 音色参考自提取法:先出一段带角色台词的视频→剪映分离音频→作为该角色音色参考传入后续生成(H3 音频参考 ≤3 段);备选 cosyvoice TTS 对轨 |
| 一段视频只有一个镜头、呆板 | 多镜头段落公式为默认(骨架 A):时间码镜头块+HARD CUT/特殊转场,单镜 1-2s 最稳;"一段一镜"是需要理由的特例 |
| 参考素材只用一两张,浪费模型能力 | 多媒体参考能用尽用:H3 支持图 ≤9/视频参考 ≤3 段/音频 ≤3/总数 ≤12——角色/道具/场景/首尾帧/上一段尾帧/视频参考/音色参考全部链接进 ACTIVE REFERENCES |
| 运镜单一、转场呆板 | 40 种电影运镜库 + 8 种转场库(生成端写法+剪辑端手法+适用时机)+ 高风险运镜双方案制 + 效果库不许落灰硬规则(运镜/转场种类数须匹配剧情复杂度)(cinematography.md) |
| 视频提示词没有章法 | 统一多镜头公式(骨架 A):SCENE CONTEXT / ACTIVE REFERENCES / CONTINUITY RULE / FORMAT / 时间码镜头块+表演注("不是X而是Y")/ ENDING FRAME / PERFORMANCE LOCK / AUDIO / LIGHTING / NO——提炼自优秀成品提示词(prompt-rules.md) |
| 长剧/多集遗忘细节 | 系列架构:00_剧情状态.md 每集强制更新(剧情摘要/角色状态/道具状态/尾帧)+ 尾帧续接机制 |
| 剧情被擅自删改 | 情节点清单逐点登记,分镜 100% 覆盖校验;矛盾模糊先问用户(GATE0 对齐义务) |
| 生成的资产用不上浪费钱 | 资产引用矩阵先行:零引用资产不生成 |
| 背景音效硬中断突兀 | 声音桥设计(段尾渐弱/延续音)+ 06 转场音频处理(J-cut/L-cut/交叉淡化) |
| 生成失败不知道怎么救 | 回退阶梯:微调重试→拆镜→保底运镜→问用户 |
Related Skills
Code Idea
"code this idea" — turns a plan or idea into a complete, AI-coding-agent-ready docs set (AGENTS.md + linked ar
Word Compiler
Context compiler for long-form fiction that turns character, voice, and scene plans into guided LLM prose gene
Venice Video Model Routing
Claude Code agent skill: Venice AI image/video generation + intelligent model routing for character consistenc
10d10s Bonus AI Ad Studio
Claude Code skill - from idea to finished AI ad with motion titles: pick the title style first, plan the Seeda
Higgsfield Prompts
Seedance 2.5 prompt skill for Higgsfield AI video: image-to-video, motion transfer, multi-shot and 30s one-tak
Higgsfield Seedance Prompt Skill
Seedance 2.5 prompt skill for Higgsfield AI video: image-to-video, motion transfer, multi-shot and 30s one-tak
Related Agents
Storyboard Writer
Storyboard writer for the video studio team. Turns vague ideas or long source material into shot-by-shot scrip
Scene Prompt Smith
Use this agent when drafting kie.ai Nano Banana 2 image prompts for the scenes of a short video. It turns a sc
Scriptwriter Storyboard
Role: You are the Scriptwriter + Storyboard agent. You turn a chosen concept into a shot-by-shot storyboard, a