ibrahimshareef96-sys

Snapcut — Communication skill for Claude Code

Communication community

An Agent Skill that edits a vertical talking-head take into a finished Reel, TikTok or Short: word-synced captions, motion graphics on the spoken word, a relit room, sound, -14 LUFS.

How to install Snapcut

This entry records only its repository, not the path inside it, so there is no exact command to give. Open ibrahimshareef96-sys/snapcut and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Snapcut does

An Agent Skill that edits a vertical talking-head take into a finished Reel, TikTok or Short: word-synced captions, motion graphics on the spoken word, a relit room, sound, -14 LUFS. Claude Code, Codex, Kimi Code.

Alternatives in Communication

  • NullMotion — Motion graphics preview and YouTube clip maker: finished ad over HyperFrames drafts, MP4 export, pause cuts, c 582 ★
  • Post Shortform — Publish a short vertical video with a caption to TikTok, Instagram Reels, and YouTube Shorts simultaneously 212 ★
  • MulmoChat — Research prototype for chat that is not a text stream: keep talking while images, maps and playable games appe 176 ★

README

snapcut

An Agent Skill that edits a raw vertical talking-head take into a finished Reel, TikTok or Short: the pauses cut, kinetic captions on the exact spoken word, motion graphics that land on the words they illustrate, the room relit with RGB streamer light, synthesised sound effects, audio mastered to -14 LUFS. Optional AI shots through Higgsfield: a restyled room, a studio cutaway, explosion plates.

You give your coding agent a clip. It transcribes it, proposes the cut and the graphics, and renders the video on your machine. Your face and voice stay exactly as filmed.

It was built from one real edit, the snap reel (a 73 s selfie take cut to 55 s, with a DM bubble, tool tiles, an AI studio shot and a finger-snap explosion). That reel's plan and beat files ship in `examples/snap-reel/` as the worked example.

Install

snapcut is a folder with a `SKILL.md` inside. Install it by cloning it where your agent looks for skills.

**Claude Code** (every project on this machine):

git clone https://github.com/ibrahimshareef96-sys/snapcut.git ~/.claude/skills/snapcut

Then type `/snapcut`, or just ask for an edit. For one project only, clone into `.claude/skills/snapcut` in that repository (use `git submodule add` if you want to commit it).

**OpenAI Codex** (CLI, IDE extension, desktop app):

git clone https://github.com/ibrahimshareef96-sys/snapcut.git ~/.agents/skills/snapcut

Then type `$snapcut`, or pick it from `/skills`.

**Kimi Code CLI**: Kimi reads `~/.claude/skills/`, so the Claude Code install above works for it too. Otherwise:

git clone https://github.com/ibrahimshareef96-sys/snapcut.git ~/.config/agents/skills/snapcut

Then type `/skill:snapcut` in a new session.

**Update** with `git -C pull`.

Requirements

  • macOS 12 or later (the person matte uses Apple Vision), with the Xcode command line tools
  • Python 3.10+, ffmpeg, curl, unzip
  • python3 -m pip install -r requirements.txt, then `python3 -m