Communitygithub.com

lzcjsyr/Book2Video

A streamlined script generation skill that directly reads books/documents, extracts core arguments, and generates high-quality, conversational video scripts for platforms like Video Account (WeChat Channels) and Douyin, outputting in standard JSON format.

What is Book2Video?

Book2Video is a minimalistic skill designed for developers and content creators using AI agent platforms like Claude Code, Cursor, or Codex. It automates the process of converting long-form documents or books into engaging, spoken-word video scripts suitable for short-video platforms such as Video Account (WeChat Channels) and Douyin. The skill streamlines reading, drafting, and polishing into three core steps: extracting key arguments and cases, writing a conversational first draft following a writing guide, and self-checking to remove AI-sounding language before packaging the final script into a strict JSON schema. It supports configurable target word counts and outputs a JSON object with titles, cover copy, golden quotes, and interaction hooks, making it directly consumable by downstream applications. This is ideal for quickly producing viral-ready book commentary content without complex project management, as all intermediate files are saved locally in the `text` folder.

Works with~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/lzcjsyr/Book2Video/tree/HEAD/skills/step1/sample_writer

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

sample_writer - Minimal Video Script Writing Assistant

This is a simplified skill for generating long-form document/book commentary video scripts. It merges the reading, writing, and polishing processes to help you quickly produce conversational, engaging Chinese scripts for Video Account (WeChat Channels) and Douyin.

Target Word Count Configuration

This configuration is the single source of truth for the final word count of the main body in this skill.

draft_min_chars: 1000 # Minimum word count for first draft final_target_chars: 1200 # Target word count for final draft

Main Workflow

You only need to follow these three core steps; no need to keep complex override logs. Please save any intermediate outputs during the process in the working directory (text folder of the project output):

  1. Quick Reading & Organizing:

    • Quickly browse the core content of the input file (focus on the beginning, table of contents, core chapters, and conclusion).
    • Extract one core question that the entire book truly attempts to answer, the author's core answer framework, and 3 key cases or arguments.
    • Save the above extracted outline as an intermediate product to _extract.txt.
  2. First Draft Writing (Conversational):

    • Following the guidelines in references/writing_guide.md, write a first draft that is well-structured, plain-spoken, and suitable for direct oral delivery, meeting the draft_min_chars word count.
    • Save the first draft as an intermediate product to _draft_v1.txt.
  3. Self-Check & Final Packaging:

    • Perform self-correction to remove unnecessary filler words, written-language pitfalls, and "AI-sounding" vocabulary, and save the corrected final draft to _draft_final.txt.
    • Generate titles, cover copy, golden quotes, interaction prompts, and sharing prompts from _draft_final.txt, and package everything into the final JSON, writing it to {output_json}.
    • Ensure the output {output_json} can be parsed by Python json.loads and fully conforms to the JSON output contract below.

Final JSON Output Contract

Output only a strict JSON object, without any Markdown markers or extra explanation.

{
  "source_name": "Work name (e.g., The Selfish Gene), without file extension",
  "video_titles": ["3 candidate video titles, non-repeating, complete information"],
  "cover_titles": ["3 main cover titles, 4 to 10 characters, short, strong, easy to scan"],
  "cover_subtitles": ["3 cover subtitles, within 15 characters"],
  "golden_quotes": ["3 opening golden quotes, strong opinion, attention-grabbing"],
  "comment_hook_options": ["3 comment prompts, prefer judgment sentences, contrast sentences, or follow-up questions; not marketing jargon"],
  "share_hook_options": ["3 sharing prompts that sound like something viewers would naturally share after watching"],
  "content": "Complete spoken final draft, around final_target_chars word count, preserving natural paragraph breaks, no Markdown markers or subtitles",
  "total_length": 0
}
  • total_length is the actual character count of content.
  • content must be completely conversational, preserve natural paragraph breaks, and absolutely must not contain subtitles, Markdown, or list formatting.

Individual skills in this repo

This repo contains 6 individual skills — each has its own dedicated page.

lzcjsyr/Book2Video

将 PDF、EPUB、DOCX 等书籍或长文档一次性转成约 4 分钟的中文读书带货短视频口播稿,以情绪共鸣、观看留存和购买欲为核心,直接输出 Book2Video 后续流程可读取的 raw JSON。用户提到读书带货、图书推广、卖书口播、商品卡、走心书单视频、情绪型读书文案或要求生成可直接进入视频流水线的带货稿时使用。

lzcjsyr/Book2Video

将长书、长文档转成中文视频号读书解说稿;先抽取文本并建立可恢复覆盖台账,覆盖自检通过后再按入口定义的正文长度生成初稿、多轮修订和最终 raw JSON。

lzcjsyr/Book2Video

将长书、长文档、PDF、EPUB、读书资料、思想经典、社科著作、商业书或研究型长文转成更有传播感的中文视频号/抖音读书口播稿,并输出项目可用的 raw JSON。只要用户提到“读书解说”“拆书稿”“视频号口播”“把这本书做成短视频”“长文档转视频脚本”“生成标题/封面/金句/raw JSON”,都应使用此 skill。

lzcjsyr/Book2Video

Cody的公众号长文写作skill。当用户需要撰写公众号文章、写稿子、续写文章、根据素材产出长文时使用。触发词包括但不限于:写文章、写稿子、帮我写、续写、扩写、公众号文章、长文、出稿、按我的风格写。即使用户只是说"帮我把这个写成文章"或"用我的风格写一下",只要上下文涉及内容创作和公众号输出,都应该触发。也适用于用户丢过来一个PDF、brief、新闻链接、语音转文字或任何素材说"帮我写篇文章"的场景。不要用于短内容(小红书帖子、推特、朋友圈)或纯标题摘要生成。

lzcjsyr/Book2Video

将中短篇专业报告、公司财报、券商研报、招股书、宏观经济分析及含有较多表格和数据的PDF文档,转化为高品质、数据精准的中文视频号/抖音口播解说稿。只要用户提到“研报”、“财报”、“商业分析报告”或需要“解读数据”、“把报告/PDF做成短视频”,务必使用此 Skill。

lzcjsyr/Book2Video

The fallback workflow for authoring custom HyperFrames video compositions at any length or format — longer or multi-scene pieces, brand / sizzle reels, montages, title cards, static loops, and freeform compositions. Input- and length-agnostic. If a specialized workflow clearly fits the input — a marketed product, a website, a topic explainer, a GitHub PR, existing footage, a short motion graphic, or a Remotion port — prefer it (see /hyperframes); use this only as the general fallback when none fit.

Related Skills