The TWO Claude Skills that run our entire AI video pipeline (free download + easy setup)
Shows a practical, low‑cost workflow that lets indie AI filmmakers produce multi‑shot videos without massive credit spend, a timely need as generation costs rise.
What this lesson covers
Joey demonstrates how he built two Claude skills that streamline prompting for Higgsfield, cutting credit usage and improving AI video output. He walks through the pipeline—from character sheets to scene generation—and shares the skills for free.
Key takeaways from the creator
AI-extracted notes, not independently verified product claims. Timestamp links let you check each point in the original video.
- 00:04 ↗
Joey built custom Claude skills to automate prompt creation for Higgsfield pipelines.
- 02:08 ↗
He notes the Higgsfield Canvas feature exists but is currently too advanced for his workflow.
- 04:00 ↗
The end‑to‑end pipeline was built in two weeks, covering character sheets, outfits, and scene prompts.
- 05:31 ↗
The two skills—Banana Pro Director and Cinema World Builder—handle image prompts, camera language, and grading automatically.
- 08:18 ↗
To use the pipeline you create a character sheet, generate a scene prompt via Claude, upload reference images, and let the skill output video parameters.
Workflow outlined in the video
- Create a detailed character sheet (appearance, outfit, personality) using the Banana Pro Director Claude skill.
- Define the desired scene (location, time of day, lens style) and feed it to the Cinema World Builder skill.
- Upload any reference images to Claude; let it expand them into full scene prompts with camera and grading specs.
- Copy the generated prompts into Higgsfield’s video generation interface, selecting appropriate cinema mode.
- Review the output, adjust credit‑heavy sections using the Cost Optimization tips, and iterate as needed.
Before you use this workflow
These notes describe the source video at its publication date. Model access, pricing, connectors and interfaces may have changed. Check the original source and the provider’s current documentation before installing an add-on, connecting an account or spending credits.
ReelStack has not independently tested this workflow. Preview one representative shot and check motion, continuity and output quality before applying it to a full production. No result, cost saving or model capability is guaranteed.
Explore the tools
How to interpret the numbers
242,687 views captured 24 October 2026. ReelStack recommendation score: 68/100. Formula: radar-v5. This is ReelStack’s calculation, not a YouTube rating or a measure of factual accuracy.
Show the calculation
- 65% performance: 8.99× current views vs the channel's recent median; same-age history is not yet available. View-evidence factor 100% (views / (views + 1,000))
- 25% freshness: 0/100 with a 14-day half-life
- Momentum pending: collecting daily snapshots; its 20% weight goes to observed performance, not free points
- 10% engagement: 100/100; likes + 4× comments, smoothed with a 500-view neutral prior