FreeSix video editing skills

Raw footage in. Your edit, your style.

Give Claude Code or Codex a practical workflow for Diffusion Studio: sync your footage, remove retakes, build motion graphics, balance audio, add captions, and check the edit.

Email me the skill pack

Free skill pack. No checkout. Uses your existing agent subscription and local tools; hardware and setup requirements apply.

Your existing agentWork with Claude Code or Codex. Agent subscription or usage costs still apply.
Local-first editingUse whisper.cpp, ffmpeg, and local subject masks instead of paid cloud analysis.
Your own styleSix SKILL.md files, scripts, and templates. Bring your own fonts, colours, and pacing.

From footage to timeline

Describe the edit. Review it on the timeline.

The pack gives your agent an editing sequence, not a fixed visual style. You direct the result and review the changes.

01

Clean up the recording

Sync recordings, transcribe locally, remove retakes and dead air, and verify the joins before building the rest of the edit.

02

Add the visual layer

Use HTML animation and three.js templates for motion graphics, b-roll, and graphics behind the speaker with local masks.

03

Finish and check

Balance voice and music, generate captions from the edited audio, and inspect frames and splices before exporting.

Inside the pack

Six skills. One editing workflow.

Start with ds-edit. It coordinates the focused skills for cutting, graphics, sound, captions, and verification.

01

ds-edit

The end-to-end playbook: briefing, cut lists, framing, layer order, and checks at every stage.

02

ds-sync-cut

Local transcription, speech islands, retake removal, verified splices, and a rebuildable edit decision list.

03

ds-motion-graphics

HTML and anime.js overlays, three.js scenes, subject masks, grain, and renderer pitfalls.

04

ds-audio

Audio-only voice tracks, loudness matching, music loops, and playback-audio syncing.

05

ds-captions

Captions from the edited voice, presets, placement, and trimming around overlays.

06

ds-qa

Structural checks, frame captures, splice re-transcription, and a final edited-voice transcript.

How it runs

Install the pack. Bring your footage.

Run Diffusion Studio with its MCP server connected to your agent. Install ffmpeg, whisper.cpp with a model, and Python 3; one audio helper also needs numpy.

diffusion-studio.md
# 1. Unzip and copy the six skill folders
~/.claude/skills/ds-*

# 2. Connect Diffusion Studio MCP to your agent
Start with ds-edit and your own style brief.

# 3. Keep analysis local
whisper.cpp → transcription
ffmpeg      → audio analysis
sam-2.1     → local subject masks

# 4. Review before exporting
check → capture → revise → approve export

Example editing brief

A clear brief. A reviewable edit.

This is an illustrative prompt, not a promised result. Give your agent a style reference and review the timeline before export.

editing-brief.md
# Illustrative prompt — adapt to your footage

START WITH ds-edit
Sync the screen recording, facecam, and separate mic.
Remove retakes and filler. Verify each splice.

USE MY STYLE
Follow my reference for fonts, colours, and pacing.
Keep motion graphics and captions clear of my face.

KEEP ANALYSIS LOCAL
Use whisper.cpp, ffmpeg, and local sam-2.1 masks.
Ask before any paid generation or cloud tools.

REVIEW FIRST
Run QA and show captures. Ask before exporting.

Original package

Six skills. No checkout.

The original archive includes six editing skills, a setup README, helper scripts, and motion-graphics templates. The tutorial transcript is not included.

  • Six SKILL.md workflow guides
  • README.md setup instructions
  • Local cutting and audio helpers
  • Motion-graphics templates
  • Caption and QA workflows
Get the Diffusion Studio pack by email

diffusion-studio-skills.zip · 6 skills · email delivery

Get the archive

Send the Diffusion Studio pack to your inbox.

Enter your details and we will email you an expiring download link. The ZIP is not published on this page.

FREE DOWNLOADDELIVERED BY EMAILNO SPAM. UNSUBSCRIBE ANYTIME

Email me the skill pack

We use these details to deliver this archive and send occasional Applied Leverage communications.

We use your email to deliver the archive. By submitting this form, you agree to receive communications from Applied Leverage. We may occasionally send marketing emails, but we promise not to spam you. Unsubscribe anytime. We do not sell your information. See our Privacy Policy and Terms.

Compatibility

Know what you need. Then start editing.

Is the whole workflow free?

The skill pack is free. The tutorial uses existing Claude Code or Codex subscriptions; agent usage is not included. The pack is designed to avoid Diffusion AI credits for editing analysis, not to remove every possible cost. Optional cloud image, video, audio generation, and other paid tools can incur charges.

What setup and hardware do I need?

The tutorial demonstrates the macOS app with Diffusion Studio MCP connected to an agent. You need ffmpeg / ffprobe, whisper.cpp and a downloaded model, Python 3, and numpy for one audio helper. Local transcription, subject masks, and rendering depend on your hardware, memory, and available storage. Setup and performance vary.

Does it include Lucas’s visual style?

No. These are style-neutral editing instructions and templates. Bring your own references, fonts, colours, and pacing. Copy the six folders into your agent’s skills directory; other agents must support SKILL.md instructions and the Diffusion Studio MCP connection. Review the scripts and prerequisites before running them.

An editing workflow handles footage. An operating loop handles recurring work.

Bring one recurring workflow. We will map the sources, review gate, output, owner, and first live cycle.

Map My First LoopSee the install blueprint