manga-illustrator

Generates consistent manga video frames, character sheets, and prompt packs from storyboard inputs.

2|1|Updated Mar 7, 2026
One-click install
npx skills add https://github.com/AutoByteus/autobyteus-agents --skill manga-illustrator-autobyteus
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: manga-illustrator
Source: https://github.com/AutoByteus/autobyteus-agents/tree/main/agent-teams/manga-video-studio-team/agents/manga-illustrator/skills/manga-illustrator
Command: npx skills add https://github.com/AutoByteus/autobyteus-agents --skill manga-illustrator-autobyteus

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Producing manga visuals for video or print requires keeping character identity, style, and aspect ratio consistent across dozens of generated images, which is hard when prompts are written ad hoc and results are not logged. ## Core Features & Use Cases - Locked Visual Language: Builds a chapter visual style guide that inherits the series canon, including the locked video aspect ratio, color strategy, and linework rules. - Character Identity Control: Generates reusable character reference sheets first and feeds them back into later image calls via input references to prevent identity drift. - Durable Prompt Pack and Logging: Writes self-contained prompts per storyboard unit with negative constraints, and logs every image-tool call with prompts, dimensions, and approval status. - Use Case: Given a storyboard for chapter 3 of a vertical 9:16 motion-comic series, produce approved full-bleed video frames, a prompt pack, and a visual production package ready for the voice and video producer. ## Quick Start Use the manga-illustrator skill to generate the full visual package for chapter 3 from its storyboard, starting with character reference sheets and ending with the visual production package.

Frequently Asked Questions about manga-illustrator

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

FAQPage Schema
How do I keep manga characters consistent across AI-generated images?▼

Generate one reusable character reference sheet per recurring character first, then feed those approved sheets back into later generate_image calls through input_images. Store approved sheets in a shared-assets folder so future chapters start from the same locked identity.

How do I generate manga frames for vertical video instead of comic pages?▼

Use a video-frame render contract where every prompt explicitly states the locked aspect ratio, such as single full-bleed 9:16 vertical motion-comic frame. Add negative prompts banning page borders, panel grids, white margins, and collages so the model does not produce a manga page layout.

Why should image generation calls run sequentially with a cooldown?▼

Sequential calls let you inspect and log each result before proceeding, catching identity drift early instead of discovering it across a whole batch. The sleep 60 cooldown after every call, including failures, prevents rate-limit issues and enforces deliberate review.

What should a self-contained image generation prompt include?▼

Include the render-unit id, contract and art state, aspect ratio and orientation in prompt text, scene and visible action, character lock details, style and color mode, text strategy, and negative constraints. Do not rely on metadata fields or generation_config to carry dimensions.

When should I use edit_image instead of regenerating an image?▼

Use edit_image when an otherwise strong image needs targeted local fixes, such as a wrong prop or small detail error. Regenerating the entire image blindly risks losing the identity and composition elements that already worked.

What are the limitations of prompt-only image generation for manga?▼

Prompt-only generation cannot reliably preserve character identity across scenes without reference images. For multi-character or recurring-location scenes, always include approved reference paths in input_images rather than hoping the prompt text alone prevents drift.