Draftcut

Turn natural speech into a refined original-voice narration and a polished video. Review the full spoken manuscript in the conversation, then automatically deliver a clean original-voice video after explicit approval. Graphic packaging is a separate stage with its own director-script approval.

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add PoetCoderJun/DraftCut --skill draftcut
Manual install
git clone --depth 1 https://github.com/PoetCoderJun/DraftCut.git /tmp/DraftCut
mkdir -p ~/.claude/skills
cp -r /tmp/DraftCut/. ~/.claude/skills/draftcut

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

From the SKILL.md

# DraftCut

自然说话 → 对话中审核完整文稿 → 批准后自动交付干净成片 → 按需审核导演脚本 → 包装成片。输入一份录音或一支口播视频,可选素材和参考片;保留用户声音,不用 TTS 补词。所有工作放入新的独立项目目录,原片不覆盖。

## 从一句请求开始

用户只需说“请使用 DraftCut 的 Skill,把这个文件做成一个视频。我想要的样式和风格是……”。以下流程、独立目录、原声保护、两道审核由本 Skill 默认负责,不要求用户补写操作步骤。Agent 自选新的项目目录;风格缺少细节时结合内容提出一个方案,不为可选偏好连续追问。已有可靠转写和健康依赖优先复用。

DraftCut 是本次专用工作流的负责人,HyperFrames 用作底层制作与渲染。按需要读取 HyperFrames 的结构、动画和 CLI 规范;没有外部 HyperFrames Skill 时,使用本仓库 references 与锁定 CLI 的 `--help`,不依赖作者安装的私有 Skill,不另启一套重复的用户创意访谈;将本次已有素材/风格写入工程 brief,后续问题集中在可读提案的必要取舍上。

直接在当前命令行/Codex 对话正文中输出完整可读稿件、简短删改理由和导演方案,不能只给文件路径或下载链接代替正文。用户以自然语言确认具体版本;不另建展示网页或操作平台,不把内部 hash、JSON、目录规划、测试签名命令当作用户的制作工作。内部证据由 Agent 管理。干净成片阶段按宿主 Agent 读取到的真实、直接用户消息确认:完整文稿已展示后,用户明确批准当前版本,就在同一任务继续生成干净成片,不再要求终端签名或额外“开始/继续”指令。含修改条件、含糊回复、历史无关批准、Agent自审或模拟用户均不算批准。工程中的对话审计记录仅供追溯,不是加密回执,不能宣称为可信宿主签名。现有签名版 HyperFrames 包装路径仍按[审批设置](docs/approvals.md)和[对话批准边界](docs/conversational-approvals.md)处理,不伪造其回执,也不把干净成片审计记录冒充为那条路径的签名。

默认以**完整文稿审核**推进:通读逐词转写,输出精炼口播全文、删改依据、纠错疑点与完整重映射字幕。试听是可选辅助,不是审核包输出、用户文稿批准或后续制作的前置条件。没有音频感知能力时记录 `hearing=not-tested` 和真实的零试听时长,按全文文字复核与时间线检查继续;不得反复要求用户听完整原片或代理,也不得虚构试听。词语疑点结合已有草稿、授权范围内的ASR结果及用户确认解决,未解决的明确标注。用户只要求文稿审核输出且明确不制作时,在输出完整文稿后停止;普通视频制作任务则在文稿获批后自动交付干净成片。用户只需要文稿时,必须把完整候选口播、简短删改理由和未解决词语直接输出到当前命令行/对话,不为展示另存一份文稿文件,不用文件路径或链接代替正文,不要求用户打开文件;内部审核包、SRT、代理与证据可按工程需要保存,不作为用户的阅读入口。

## 明确授权的模拟用户测试
Read the full skill on GitHub

Details

Skill name
draftcut
Repo updated
21 October 2026

FAQ

What does the Draftcut skill do?

Turn natural speech into a refined original-voice narration and a polished video. Review the full spoken manuscript in the conversation, then automatically deliver a clean original-voice video after explicit approval. Graphic packaging is a separate stage with its own director-script approval.

How do I install the Draftcut skill?

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

Is Draftcut popular?

It lives in PoetCoderJun/DraftCut, which has 12 GitHub stars and 0 forks.

Related skills

Genpark Snap Sell

12

alphaparkinc/genpark-snap-sell

Sell anything with a single photo. An autonomous creator monetization agent that turns an image into a live, zero-fee Stripe checkout page in seconds.

Images, Video & Audio

Deslop

12

skyzer/deslop-the-copy

Deslop, humanize, edit, or audit AI-sounding, robotic, generic, corporate, or ChatGPT-like copy while preserving the writer's voice.

Images, Video & Audio

Gpt Images

12

lukasersil/gpt-images

Turns a plain-language brief into a production-ready prompt for OpenAI GPT Image models (ChatGPT Images 2.5, gpt-image-2.5-flare, gpt-image-2.5-sunburst, gpt-image-2). Returns a seven-block prompt, a settings block for the chosen surface, and three concrete levers for the next iteration. Covers text inside the image, reference binding, edits that do not drift, transparent backgrounds, and the hard API limits on size, quality and input images. Triggers: 'write an image prompt', 'gpt image prompt', 'chatgpt images 2.5', 'make me a poster', 'product shot prompt', 'article cover image', 'edit this

Images, Video & Audio

Chatcut to Fcpxml

10

AFangTeam/chatcut-to-fcpxml

Clone a ChatCut project into Final Cut Pro via FCPXML with frame accuracy — cuts, multi-track overlays, zoom moves (static punch-ins and keyframed slow pushes), music/SFX with volumes and fades, and native iTT captions. Use this skill whenever a user wants to export, hand off, convert, or "open in Final Cut Pro / FCP" a ChatCut project or timeline, wants an .fcpxml file, or complains that an FCPXML import into FCP dropped media, looks wrong, or lost zooms/captions. Also use it to verify an FCPXML import inside Final Cut Pro.

Images, Video & Audio

Image Audit

12

xiaowu89/skill-function

对图片进行鉴黄、政治、暴恐内容审核。先将图片压缩到 500px/JPEG 后直传 NX API 审核,以表格汇总结果。适用于用户提到图片审核、内容检查、鉴黄、政治识别、暴恐识别、违规扫描、图片安全、JPG/PNG/WebP 审核的场景。

Images, Video & Audio

Screenstudio Alt

11

connerkward/screenstudio-alt-skill

Screen-Studio-style post-production for screen recordings, headless — auto speed-up of idle, auto-zoom on click clusters, keystroke overlay chips, smoothed synthetic cursor, and 9:16 vertical export that follows the action. Use when polishing a screen recording / demo video for sharing, when the user mentions Screen Studio, auto-zoom, idle speed-up, or vertical/social video from a screen capture, and for any social-facing demo (vertical output is the default for those).

Images, Video & Audio

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