AI repository intelligence

FR-AI / ANALYSIS

Why it is worth attention

It systematically bridges reference video analysis and component-based video generation, with a five-level fidelity QA process, making it a rare practical resource for AI-driven motion design rather than static slide production.

Who it is for

AI video creators and motion designers
Developers using Codex, Claude Code, or Cursor for video generation
SaaS/AI product marketers needing style-consistent shorts
Video quality engineers looking for reproducible replication workflows

Use cases

Planning and generating a dark SaaS product short with motion-first direction
Replicating a reference video effect with pixel-level or visual-level fidelity and component extraction
Producing black-white-text animated opener with synchronized typing sound for video intros
Decomposing a reference video into parameterized motion components for future reuse

Strengths

Provides a complete workflow from theme to motion thesis, beat graph, component scheduling, and anti-PPT quality gate
Includes a reference-video replication skill with five-level fidelity classification (source, pixel, frame, visual, style) and multiple evidence metrics
Designed for integration with popular AI coding agents (Codex, Claude Code, Cursor) via Skills CLI
Long-term maintenance promise with planned expansion across video types and production stages

Considerations

The skills are not standalone rendering tools; they require HyperFrames/Remotion and asset pipelines
Each skill has explicit design boundaries (e.g., ai-motion-director does not handle final rendering, subtitles, or audio)
Adoption requires familiarity with the Skills CLI and the specific agent environments; not a plug-and-play solution for non-technical users

README quick start

Video Production Skills

这是一个长期维护的 AI 视频制作 skill 精选仓库，用来沉淀我在视频创作、视频复刻、动效生成、片头包装、质检交付等流程中反复使用的方法，也用来把优秀参考视频中的动效拆成 HyperFrames / Remotion 能够实现和复用的组件。

仓库不是只服务于某一个视频项目。后续优秀的视频制作 skill 都会优先收录在这里，并按“视频类型 / 风格 / 工作流环节”持续扩展。当前收录的是 AI 动效导演元 Skill、参考视频复刻质检、暗色 SaaS / AI 产品短片、黑底白字打字开场。

这些 skill 适用于 Codex、Claude Code、Cursor 等支持本地 skill 目录或 Skills CLI 的 AI 编程/创作代理。仓库中的每个一级目录都是一个独立可安装的 skill。

长期目标：从复刻到组件库

reference-video-replica-qc 的意义不是单纯复制某一个参考视频，而是把参考视频里真正让画面“像视频”的东西拆出来：镜头节奏、入场方式、路径运动、组件状态切换、字幕/标题层级、模型列表、线条轨道、卡片翻转、光效、遮罩、缩放和转场。每一次复刻都应该把这些动效整理成 HyperFrames / Remotion / React / CSS / SVG / GSAP 可以编写、组合和参数化的技术组件。

长期来看，这个仓库会形成一个视频动效组件库。每个通过复刻验证的组件都要留下描述：它适合什么内容、解决什么画面问题、需要哪些输入参数、时间线如何运动、应该用在哪类视频段落、有什么对齐证据和限制。以后 AI 制作视频时，就不只是生成一页页静态版式，而是能够根据脚本内容自由选择这些组件，把开场、解释、列表、对比、流程、CTA、产品展示等段落组合成更接近真实视频语言的成片。

ai-motion-director 是这套体系的元导演层：给它一个主题、脚本或简报，它先决定视频的运动隐喻、beat graph、组件调度、素材需求和反 PPT 质检标准，再把任务交给 HyperFrames / Remotion / 生图 / 具体视频 skill 执行。

已收录 Skills

Skill	视频制作环节	适合做什么	视频风格 / 方法	详情
`ai-motion-director`AI 动效导演元 Skill	主题到动效语言 / 元导演 / 反 PPT 质检	给一个主题、脚本或简报，先建立 motion thesis、beat graph、组件/素材计划，再控制 HyperFrames/Remotion 进入制作	Motion-first：视觉隐喻、连续时间线、状态变化、运动语法、anti-PPT gate	查看页面
`reference-video-replica-qc`	复刻闭环 / 组件沉淀 / 质检	参考视频拆解、动效复刻、HyperFrames/Remotion 组件化、五级保真度判定	先采样定位、再全帧验收；素材门、运行时门、交付门；MAE/时间偏移/边界帧/PSNR/SSIM/哈希证据	查看页面
`dark-saas-magic-video`暗色 SaaS 魔术短片	风格化正片生成	暗色 SaaS / AI 产品短片、工具发布视频、产品能力展示	Presenton-like magic UI：黑色空间、底部紫光、动感大字、渐变 CTA、漂浮 UI、模型环、导出物件	查看页面
`black-white-text-opener`	片头包装	新视频片头、教程开场、观点视频开头、产品视频引子	纯黑/近黑背景 + 白色大字逐字打出 + 同步 typing click 音效 + 干净转场	查看页面

效果预览

AI 动效导演元 Skill

这是流程控制 skill，暂无单独预览视频；它用于在制作前决定运动隐喻、beat graph、组件调度和反 PPT 质检标准，并控制后续视频 skill 产出成片。

黑底白字打字开场

▶ Watch Black White Text Opener

暗色 SaaS / Magic UI

示例成片：雪踏乌云暗色 SaaS 介绍短片

▶ Watch Xuetawuyun Dark SaaS Showcase

参考视频复刻闭环

示例成片：Presenton 复刻 Bitexact 成片

[▶ Watch Presenton Replica Bitexact Showcase](https://github.com/user-attachments/assets/f792bd12-d8b3-43b7-b751-98aeb03371

Description

Reusable AI video production skills library for creation, recreation, motion design, openers, and QA.

Related repositories

Similar projects matched by category, topics, and programming language.

lopopolo

Featured

harness-engineering

Harness Engineering is a methodology for improving coding agent outputs by carefully crafting the environment around them—providing curated context, tools, and executable constraints that encode an organization’s nonfunctional requirements and cumulative lessons.

AI & Machine LearningAI Agents

2,390

slvDev

Featured

esp32-ai

A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.

AI & Machine LearningLarge Language Models

1,960

littledivy

Featured

mimic

mimic captures traffic from any iOS or web app and automatically generates a Python client library that lets you call the app's API like a regular library.

AI & Machine Learning

1,482