Video Reverse Engineering

逆向拆解一条视频的拍摄与制作方法：给定一个视频链接，下载并抽取关键帧、字幕与音频，产出(1)分镜脚本、(2)可直接用于AI绘图/生图工具的制图提示词（含图生视频运镜提示）、(3)完整复刻/仿拍制作指南。当用户丢来一个视频网址（YouTube、B站/Bilibili、抖音、快手、小红书、TikTok、Instagram Reels等）并表达"分析这个视频""这个视频怎么拍/怎么做的""拆解一下这个视频""复刻/仿拍这个视频""学习这个视频的运镜和风格"等意图，或英文 analyze/breakdown/reverse-engineer/recreate this video 之类请求时，都要主动使用本Skill——即便用户只发了一个链接、没有说"拆解"两个字，只要意图是理解或复刻视频的制作方法而不是单纯想要内容摘要，就应触发。

5 starsImages, Video & Audio

Install

skills CLI (Claude Code, Cursor, Codex and more)

npx skills add whaojie797-design/video-reverse-engineering --skill video-reverse-engineering

Manual install

git clone --depth 1 https://github.com/whaojie797-design/video-reverse-engineering.git /tmp/video-reverse-engineering
mkdir -p ~/.claude/skills
cp -r /tmp/video-reverse-engineering/. ~/.claude/skills/video-reverse-engineering

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# 视频逆向拆解与复刻

## 这个 Skill 在做什么，以及为什么这样设计

用户甩来一个视频链接，目的不是"看懂这条视频讲了什么内容"（那是普通的视频总结），而是"看懂这条视频
是怎么被做出来的"，然后能照着做一条类似的。这两件事需要的分析深度完全不同：内容总结关心信息，这个
Skill 关心**制作层面**——镜头怎么分、怎么运镜、光线怎么打、什么调色风格、剪辑节奏多快、每一帧要用
什么样的 AI 生图提示词才能重新画出来。

所以整个流程不是"看一遍视频凭印象写感想"，而是先把视频拆成真实存在的关键帧（用场景检测算法抽出来，
而不是脑补），逐帧做视觉分析，再综合成三份产出物。这样产出的分镜脚本时间码是精确到秒/帧的，制图
提示词是能直接丢进 Midjourney / 即梦 / 可灵 这类工具用的，而不是"一个咖啡店的镜头"这种空泛描述。

## 触发场景

用户发来视频链接并表达出"看懂/学会这条视频制作方法"的意图，哪怕只是甩一个链接加一句"这个怎么拍
的"、或者干脆只有链接没多说什么。不要把这类请求当成"帮我总结一下这个视频在讲什么"来处理——除非用户
明确说自己只是想要内容摘要而不关心制作方法。

## 环境依赖

本 Skill 需要一个**能执行 bash 脚本、能调用 yt-dlp 与 ffmpeg、并且能查看图片文件**的运行环境，
同时网络要能访问目标视频平台。具体依赖：

- **yt-dlp**：下载视频、拉取平台字幕/元数据。
  - Linux / macOS：`pip install yt-dlp --break-system-packages`
  - Windows：`pip install yt-dlp`（不需要 `--break-system-packages`）
- **ffmpeg / ffprobe**：抽帧、抽音频、读元数据。
  - macOS：`brew install ffmpeg`
  - Linux：`apt install ffmpeg`
  - Windows：`winget install Gyan.FFmpeg` 或 `scoop install ffmpeg`
- **网络能访问目标视频平台**（YouTube/B站/抖音/小红书等）。如果当前环境的网络策略屏蔽了这些域名，
  下载会失败——处理方式见下面"素材获取失败时怎么办"。
- **读图能力**：视觉分析靠的是宿主自带的多模态读图能力，打开抽出来的关键帧逐张看，不需要额外的
  视觉模型。如果宿主不支持读图，这个 Skill 无法给出可信的镜头分析——不要凭文件名或时长瞎编画面。

## 整体工作流程

### 第一步：确认目标

默认走完下面六步，产出完整三份交付物。如果用户措辞明显只要其中一部分（比如"只要分镜脚本"），就只
产出那部分，但抽帧和逐帧分析这两步仍然要做——它们是所有产出的地基，不能跳过直接凭空写。

如果视频较长（超过 3 分钟），开工前跟用户确认一下：整支都拆，还是只拆某个片段（比如广告/短视频的
黄金开场几秒，分析价值往往远高于中后段）。视频越长逐帧分析工作量越大，没必要死磕全片。

### 第二步：抓取素材

用 `scripts/extract_shots.sh` 一次性搞定下载、抽帧、抽字幕、抽音频：

```bash
bash scripts/extract_shots.sh -u "<视频链接>" -o ./analysis
```

跑完会得到：

Read the full skill on GitHub

Details

Skill name: video-reverse-engineering
Repository: whaojie797-design/video-reverse-engineering
License: MIT
Repo updated: 5 August 2026

FAQ

What does the Video Reverse Engineering skill do?

How do I install the Video Reverse Engineering skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add whaojie797-design/video-reverse-engineering --skill video-reverse-engineering. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Video Reverse Engineering popular?

It lives in whaojie797-design/video-reverse-engineering, which has 5 GitHub stars and 0 forks.

Related skills

Xiaomimimo Extended

hecaodong/xiaomimimo-extended

Xiaomi MiMo API toolkit: web search, image / audio / video understanding, and TTS speech synthesis. 调用小米 MiMo API 的隐藏能力 —— 联网搜索、图片识别、音频转写、视频分析、语音合成(TTS)。 Trigger phrases — 触发词: EN: mimo search, web search, image analysis, screenshot OCR, audio transcription, meeting summary, video analysis, tutorial extract, text to speech, voice clone 中: mimo 搜一下 / 联网搜 / 看看这张图 / 这段录音说了啥 / 转写 / 提取视频步骤 / 合成语音 / 念出来 / 克隆音色

Images, Video & Audio

Video GIF Studio

miles990/video-gif-studio

Plan scripts and storyboards, analyze music cue points, create character animation and game sprite exports, or edit existing video. Use for GIF, APNG, transparent frames, sprite sheets and video deliverables from text, images, video or music. Grok is the default motion source; other available video tools, image keyframes and local media are alternatives.

Images, Video & Audio

Yc Application Coach

evalyze-ai/yc-application-coach

Helps founders write, review, pressure-test and strengthen a Y Combinator (YC) application, record a good founder video, and prepare for the YC interview, grounded in YC's own pages, partner statements and real accepted and rejected applications. Use whenever someone mentions applying to YC or Y Combinator, any YC application question (the 50-character description, what are you making, how far along, who writes code, competitors, how will you make money, other ideas, founder profiles), the founder video or its transcript, the demo, a YC interview invite or mock interview, a YC rejection, reapp

Images, Video & Audio

Video Debug

vladzima/video-debug

Use when the user references a video, screencast, screen recording, or any path ending in .mp4, .mov, .webm, .mkv, or .gif — especially when asking to debug a UI bug, visual glitch, or reproduction recording. Extracts key frames via ffmpeg scene detection so the agent can visually inspect what went wrong.

Images, Video & Audio

Youtube to Case Study

tstfully/youtube-to-case-study

Turn any YouTube video (founder interview, product teardown, conference talk, marketing breakdown, UX walkthrough, podcast) into a beautiful case-study slideshow: one HTML file that opens in any browser, built on reveal.js, with speaker notes and every fact linked to the exact second in the video. Use this skill whenever someone wants a case study, business case study, marketing case study, UX case study, product teardown, growth teardown, competitor breakdown, swipe file, lesson deck, or presentation / slides / deck from a YouTube video or link. Also trigger on 'turn this video into a case st

Images, Video & Audio

Tiktok Ugc Seedance

xigua0626/tiktok-ugc-seedance

Convert a target ecommerce product and one TikTok competitor video into a TikTok Shop UGC strategy package: product positioning, category/search keywords for finding competitor videos, competitor video transcript and conversion-structure breakdown, 15-second English UGC voiceover script, visual shot plan, and Seedance 2.0 prompt. Use when the user provides a product link/image plus a TikTok reference link, asks to analyze a competitor TikTok video, write UGC TikTok scripts, or turn the script into Seedance prompts for cross-border ecommerce ads.

Images, Video & Audio

Browse the most popular Claude skills, all images, video & audio skills, or write your own with the skills generator.