scroll-world

Generates scroll-scrubbed camera-flight landing pages from AI-rendered scene stills and video clips.

Updated Aug 10, 2026
One-click install
npx skills add https://github.com/KryptaMedina7/quero-s-home --skill scroll-world-kryptamedina7
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: scroll-world
Source: https://github.com/KryptaMedina7/quero-s-home/tree/main/.agents/skills/scroll-world
Command: npx skills add https://github.com/KryptaMedina7/quero-s-home --skill scroll-world-kryptamedina7

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Building an immersive scroll-driven landing page where the camera flies continuously through a brand's world requires coordinating AI image generation, frame-locked video chaining, encoding, and a portable scrub engine — a pipeline that is easy to get wrong, especially at clip seams. ## Core Features & Use Cases - Guided interview and brand intake: collects subject, palette, art direction, camera style, journey sections, mobile opt-in, and budget before any rendering spend. - Seamless video chain generation: renders scene stills with Higgsfield or Codex, then chains dive and connector clips via Monid (Seedance 2.0) or Higgsfield credits with frame-identical seam handoffs. - Portable scrub engine: ships a framework-agnostic vanilla JS engine that scrubs pre-rendered video by scroll position, with mobile 9:16 variants, reduced-motion support, and iOS hardening. - Use Case: A bubble tea brand wants a homepage where scrolling flies the visitor from tea farms through the kitchen to the flagship shop; the skill generates the stills, the connected camera clips, and the wired landing page. ## Quick Start Ask the assistant to build a scroll-through-the-world landing page for your business using the scroll-world skill and answer its interview questions about subject, style, and budget.

Frequently Asked Questions about scroll-world

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a scroll-scrubbed video landing page?

The skill interviews you for subject, brand palette, art direction, and camera style, then generates scene stills, chains dive and connector video clips with frame-identical seams, and wires a vanilla JS engine that scrubs the clips by scroll position.

What video models support seamless clip chaining?

Only models accepting start-image and end-image conditioning can frame-lock seams: seedance_2_0, seedance_2_0_mini, and kling3_0 on Higgsfield, or Seedance 2.0 via Monid. Reference-only models cannot hold a seam and are declined.

Does the scroll engine work with Next.js or plain HTML?

Yes. The scrub engine is self-contained vanilla JS that builds its own DOM and injects namespaced CSS into any container, so it drops into plain HTML, Next.js, Vue, or server-rendered pages without framework dependencies.

Why do seams pop between AI-generated video clips?

Pops happen when a connector ends on a freshly rendered still instead of the neighbouring clip's actual rendered frame. The fix is extracting first and last frames from the rendered videos with ffmpeg and using those as the connector endpoints.

How much does rendering a scroll-world chain cost?

On Monid, a 1080p six-scene chain costs roughly $27 pay-per-clip, about $11 at 720p, with 480p previz around $0.28-$0.35 per clip. Higgsfield credits are the fallback biller, and the skill states the estimate before rendering.

Can I get a mobile version of the scroll experience?

Yes, as an opt-in. The skill renders a parallel native 9:16 portrait chain with its own frame-locked seams and wires clipMobile, connectorsMobile, and stillMobile variants, roughly doubling video generation cost.