bozhouDev GitHub avatar

video-skills-toolkit

bozhouDev

A modular agent skills toolkit for short video creation on Chinese social media, automating viral analysis, content remixing, voiceover, subtitles, script, animation, music, and cover generation.

Stars

106

7-day growth

No data

Forks

43

Open issues

0

License

MIT

Last updated

2026-07-27

AI repository intelligence
FR-AI / ANALYSIS

Why it is worth attention

It provides a structured, end-to-end pipeline from trend research to final polished video, integrating multiple AI services (MiniMax, ElevenLabs, Volcengine) with deterministic viral detection and clear skill boundaries.

Who it is for

  • Chinese social media content creators (Douyin, Xiaohongshu)
  • Video production teams seeking automated workflows
  • AI agent developers building content creation tools
  • Indie creators wanting to replicate viral strategies

Use cases

  • Analyze trending Douyin/Xiaohongshu videos and extract viral benchmarks
  • Remix popular video content into original drafts for different platforms
  • Generate voiceovers with MiniMax and export subtitles in multiple formats
  • Create animated talking-head videos with HyperFrames and add background music

Strengths

  • Modular skill architecture with clear boundaries and explicit dependencies
  • Uses deterministic rules for viral detection, avoiding subjective bias
  • Supports multiple platforms (Douyin, Xiaohongshu, Bilibili, YouTube) with platform-specific helpers
  • Provides test scripts and verification workflows to ensure pipeline integrity

Considerations

  • Requires multiple external API keys (MiniMax, ElevenLabs, Volcengine, etc.) and cloud storage credentials
  • Heavy runtime dependencies (Node.js, Python, FFmpeg, Playwright, Chrome) increase setup complexity
  • Personalized for a specific creator (bozhou); other users must replace reference images and writing style skills

README quick start

Video Skills Toolkit

这是一套面向自媒体视频创作的 Agent Skills 工具包。当前版本围绕“爆款调研与转写 → 内容二创 → 配音与字幕 → 导演稿 → HyperFrames 口播成片 → BGM 与抖音封面”重新组织,不再包含旧的 talking-head-remotionsketch-story-remotion

Ten Core Skills

SkillResponsibility
viral-video-benchmark扫描抖音/小红书账号近期作品,用确定性规则判断爆款,建立证据包,拆解并归档。
media-to-transcript下载平台视频,抽取与切分音频,通过火山引擎 AUC ASR 产出原始转写、纠错提示和最终逐字稿交接。
content-remix把已完成八部分拆解的爆款机制迁移成泊舟自己的小红书图文、抖音口播或小红书视频内部草稿。
minimax-voice-director为锁定逐字稿制作可审批的 MiniMax 声音导演稿,生成、选 Take 并发布最终人声。
audio-to-subtitles把最终音频/视频转成 SRT、VTT 和对齐 JSON;也保留轻量 MiniMax TTS 兼容入口。
video-script使用锁定音频和字幕生成管线无关的导演契约、Beat Graph、制作规格和 motion contract。
talking-head-hyperframes生成并验证 HyperFrames 口播项目的固定舞台、PIP、输入归档、manifest 与执行交接。
hyperframes-scene-animator实现逐镜静态页、语义动效、转场、声音 cue、proof 和预览审查闭环。
music通过 ElevenLabs Music API 生成视频 BGM,优先复用已登记音轨,并管理原始产物、无缝循环、哈希和审批状态。
douyin-cover诊断、生成和改版抖音、视频号、小红书等平台的短视频封面。

Workflow

调研:viral-video-benchmark
        └─ media-to-transcript → 最终逐字稿/关键帧 → 爆款证据与八部分拆解
        └─ content-remix → 二创迁移卡 → 已确认创作逐字稿

制作:已确认创作逐字稿
        → minimax-voice-director
        → audio-to-subtitles
        → video-script
        → talking-head-hyperframes
        → hyperframes-scene-animator
        → music(BGM)
        → douyin-cover

关键边界:

  • viral-video-benchmark 负责找对标、判断、证据和拆解,不自动生成二创成稿。
  • media-to-transcript 处理对标素材的下载、AUC ASR 和纠错;原始 ASR 不是最终逐字稿。
  • content-remix 只消费已选定、已拆解的样本,必须替换原作者事实与表达;它生成迁移方案或内部草稿,不自动发布。
  • 正式配音前必须有已确认的逐字稿;本工具包不把“边写稿边做分镜”当作正式流程。
  • minimax-voice-director 产出并审批最终人声;audio-to-subtitles 只从定稿音频建立时间轴。
  • video-script 只做导演规划,不写场景代码。
  • talking-head-hyperframes 只建固定舞台与执行交接;真正的镜头实现属于 hyperframes-scene-animator
  • music 只在已登记音轨不适用时消耗 ElevenLabs 额度;口播 BGM 默认是短、纯音乐、可循环且不抢人声。
  • douyin-cover 在标题与发布方向锁定后执行,可与成片收尾并行。

Bundled Support Skills

下列 3 个目录随仓库一起安装,但不另算创作核心:

SkillUsed byPurpose
video-transcriptmedia-to-transcript已废弃的转写入口;仅保留 B 站、抖音、小红书和 YouTube 下载/探测 helper。
[hook-writing](skills

Description

Video skills toolkit for Remotion talking-head, sketch story, and audio-to-subtitles workflows.

Related repositories

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

Vincentwei1021
Vincentwei1021 GitHub avatar

video-shotcraft

An AI agent skill that turns Claude Code or Codex into a motion-design studio for crafting cinematic product videos with Remotion, offering 106 shot recipe cards, 162 styles, 161 motion previews, and a production-ready template.

TypeScript
2,432
TanStack
Featured
TanStack GitHub avatar

router

TanStack Router is a type-safe, data-driven React router with built-in caching, prefetching, and nested layouts, while TanStack Start extends it into a full-stack SSR framework.

Web DevelopmentFrontend Frameworks
14,861
vercel-labs
Featured
vercel-labs GitHub avatar

scriptc

scriptc compiles ordinary TypeScript into small, fast native executables without needing Node.js, V8, or any JavaScript runtime in the binary.

Developer ToolsCode Quality & Build
1,985