art-design-skill

Generates character, scene, and prop text-to-image prompts for short-drama production.

Updated Aug 12, 2026
One-click install
npx skills add https://github.com/VANTSU-1874/woods-video-studio --skill art-design-skill-vantsu-1874
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: art-design-skill
Source: https://github.com/VANTSU-1874/woods-video-studio/tree/main/template/.agents/skills/art-design-skill
Command: npx skills add https://github.com/VANTSU-1874/woods-video-studio --skill art-design-skill-vantsu-1874

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Short-drama production needs consistent visual reference images for every character, scene, and prop, but writing detailed, style-consistent text-to-image prompts for an entire series by hand is slow and error-prone. This Skill converts the finalized asset bible into production-ready prompts in one batch. ## Core Features & Use Cases - Full-series asset prompt design: Reads final-asset-bible.json and the asset registry to design prompts for every character, scene, and prop variant before directing begins. - Standardized output layouts: Enforces fixed reference-image formats, such as a 16:9 character sheet with a strict front-facing facial close-up plus front/side/back three-view, and a 2x2 four-direction grid for scenes. - Variant-based prompt writing: Uses the first variant as a baseline and writes image-to-image style delta prompts for later variants, keeping core features consistent. - Validated batch upsert: Writes all prompts through a checked JSON batch into the shared prompt files and asset registry, then runs the design-stage gate. - Use Case: Given a finalized asset bible for episode 1, produce complete Chinese narrative prompts for all characters, scenes, and props, register them in asset-registry.json, and pass the design gate. ## Quick Start Ask the agent to run the art design step to generate all character, scene, and prop prompts from the finalized asset bible and upsert them into the asset registry.

Frequently Asked Questions about art-design-skill

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

FAQPage Schema
How do I generate character reference image prompts for a short drama?

Run this Skill after the asset bible is finalized. It reads final-asset-bible.json and writes a 16:9 prompt per character variant with a strict front-facing facial close-up on the left and front, side, and back full-body views on the right.

What image format do scene prompts use in this workflow?

Each scene gets one 2x2 four-grid image showing front, back, left, and right views of the same location with no people. Scene variants are only created when the script requires a different time, lighting, or weather state.

Can I edit the prompt markdown files directly?

No. The shared files for character, scene, and prop prompts plus asset-registry.json must be updated through the ly.js asset-prompts check-batch and upsert-batch commands using a validated JSON batch file.

How are multiple variants of the same character handled?

The first variant is a complete baseline prompt. Later variants start with a reference to the baseline, keep all core features unchanged, and describe only the differences such as costume, age stage, or damage state.

What happens if the batch precheck fails?

If check-batch returns ok:false, you must allocate a new tmp path, rewrite the entire JSON batch, and run the precheck again. Upserting a malformed batch is forbidden, and the design gate must pass before directing continues.