Install in seconds
Install this skill
Copy the command and run it in your terminal. You can review the source before installing.
terminal
git clone https://github.com/revfactory/harness-100

Works with Git. The repository opens in your current directory.

🎨
AI Engineering

Comic Creator

by revfactory

A multi-agent pipeline for creating comics from storyboard to final layout. Use for 4-panel strips, short-form, long-form, or webtoon production.

1.2K stars419 forksAdded 2026/07/20

Documentation

README

Comic Creator — Full Comic Production Pipeline

Produce comics from storyboard through dialogue, image generation, and editing via an agent team, all in one pass. Supports 4-panel strips, short-form, and long-form (chapter-based) formats.

Execution Mode

Agent Team — Five agents communicate directly via SendMessage and cross-validate each other's work.

Agent Roster

Agent File Role Type
storyboarder .claude/agents/storyboarder.md Synopsis, storyboard, panel layout design general-purpose
dialogue-writer .claude/agents/dialogue-writer.md Character dialogue, sound effects, narration general-purpose
image-generator .claude/agents/image-generator.md Gemini-based panel image generation general-purpose
comic-editor .claude/agents/comic-editor.md Speech bubble placement, page editing spec general-purpose
quality-reviewer .claude/agents/quality-reviewer.md Consistency validation, continuity checks general-purpose

Workflow

Phase 1: Preparation (Performed Directly by the Orchestrator)

  1. Extract from user input:
    • Topic/Story Idea: The story to turn into a comic
    • Format: 4-panel / Short-form (8-16 pages) / Long-form (chapter-based) / Webtoon
    • Genre: Comedy, drama, fantasy, slice-of-life, action, etc.
    • Art Style (optional): Manga, American comics, webtoon style, etc.
    • Existing Files (optional): Scenario, character settings, etc.
  2. Create _workspace/ directory and _workspace/panels/ subdirectory
  3. Organize input and save to _workspace/00_input.md
  4. If existing files are available, copy to _workspace/ and skip corresponding phase
  5. Determine execution mode based on request scope

Phase 2: Team Assembly and Execution

Order Task Owner Dependency Deliverable
1 Storyboard storyboarder None _workspace/01_storyboard.md
2a Dialogue writing dialogue-writer Task 1 _workspace/02_dialogue.md
2b Panel image generation image-generator Task 1 _workspace/03_image_prompts.md, _workspace/panels/
3 Page editing spec comic-editor Tasks 1, 2a, 2b _workspace/04_layout.md
4 Quality review quality-reviewer Tasks 2a, 2b, 3 _workspace/05_review_report.md

Tasks 2a (dialogue) and 2b (images) run in parallel. Both depend only on Task 1 (storyboard).

Inter-agent communication flow:

  • storyboarder completes -> delivers per-panel situations and emotions to dialogue-writer; character sheets and compositions to image-generator
  • dialogue-writer completes -> delivers SFX style to image-generator; bubble types and reading order to comic-editor
  • image-generator completes -> delivers image files and blank area info to comic-editor
  • comic-editor completes -> delivers editing spec to quality-reviewer
  • quality-reviewer cross-validates all deliverables. On RED Must Fix findings, sends revision requests -> rework -> re-validate (up to 2 cycles)

Phase 3: Integration and Final Deliverables

  1. Verify all files in _workspace/
  2. Confirm all RED Must Fix items have been addressed
  3. Report final summary to user

Execution Modes by Scope

User Request Pattern Execution Mode Agents Deployed
"Create a comic," "full production" Full Pipeline All 5 agents
"Just create a storyboard" Storyboard Mode storyboarder + reviewer
"Make a comic from this scenario" (existing file) Image Mode image-generator + comic-editor + reviewer
"Just write the dialogue" Dialogue Mode storyboarder + dialogue-writer + reviewer
"Review this comic" Review Mode reviewer only

Using Existing Files: If the user provides a scenario, storyboard, etc., copy to _workspace/ and skip that agent's phase.

Data Transfer Protocol

Strategy Method Use Case
File-based _workspace/ directory Store and share major deliverables
Message-based SendMessage Real-time key information delivery, revision requests
Task-based TaskCreate/TaskUpdate Progress tracking, dependency management

File naming convention: {order_number}_{deliverable}.{extension}

Error Handling

Error Type Strategy
Image generation failure Proceed with text prompts only; user can retry with the prompts
Character consistency breaks Strengthen character descriptions in prompts and regenerate (up to 2 times)
Agent failure Retry once -> if still failing, proceed without that deliverable; note in review report
RED found in review Send revision request -> rework -> re-validate (up to 2 cycles)
Content rejected as inappropriate Modify prompt and retry; note in report

Test Scenarios

Normal Flow

Prompt: "Create a 4-panel comic about a programmer's daily life. Comedy genre." Expected Result:

  • Storyboard: 4-panel setup-development-twist-punchline storyboard, character sheet
  • Dialogue: Character-specific speech, sound effects, punchline
  • Images: 4 panel images generated (consistent art style)
  • Editing: Speech bubble placement, reading flow specification
  • Review: Full consistency matrix verified

Existing File Flow

Prompt: "Create a webtoon from this scenario" + attached scenario file Expected Result:

  • Incorporate existing scenario into _workspace/01_storyboard.md
  • Storyboarder adds panel layout design on top of existing scenario
  • Deploy dialogue-writer + image-generator + comic-editor + reviewer

Error Flow

Prompt: "Just create a quick storyboard, any topic" Expected Result:

  • Switch to storyboard mode (storyboarder + reviewer)
  • Topic is unclear, so storyboarder proposes 3 genre-based synopses
  • Review report notes "Dialogue/image/editing not generated"

Agent Extension Skills

Agent Extension Skill Knowledge Provided
storyboarder, image-generator /panel-composition Camera angles, gaze flow, page rhythm, prompt structure
dialogue-writer, comic-editor /visual-narrative Speech bubble systems, SFX typography, Show-Don't-Tell, scene transitions
storyboarder, image-generator /character-design-system Character sheets, silhouette test, expression charts, AI consistency strategy

More from revfactory

Other Claude Code skills by this author in the directory.

🪝
1w ago

Hook Writing

Provides 15 hook patterns, viewer retention psychology, and click-to-watch conversion formulas for designing YouTube video openings that minimize drop-off. Used by scriptwriter agents or standalone for crafting compelling hooks.
Writing
+0%1.2K419
🎨
1w ago

Thumbnail Psychology

A visual psychology guide for designing YouTube thumbnails that boost click-through rate. Covers color theory, composition, text overlay rules, emotional expressions, and mobile readability to optimize thumbnails for higher CTR.
AI Engineering
+0%1.2K419
🎬
1w ago

YouTube Production

A full production pipeline where an agent team collaborates to generate YouTube video content — strategy, script, thumbnail, and SEO metadata. Use for requests like planning a YouTube video, writing a video script, creating a thumbnail, or optimizing SEO for existing content.
AI Engineering
+0%1.2K419
🎙️
1w ago

Audio Storytelling

Provides narrative structures, pacing, and sound design methodologies for creating immersive audio-only experiences. Used by podcast scriptwriters and show note editors to maximize listener retention.
Writing
+0%1.2K419
🎙️
1w ago

Interview Techniques

Provides question design, conversation flow management, and emotional arc design methodologies for drawing out deep conversations in guest interviews for podcast scripts.
Writing
+0%1.2K419
📈
1w ago

Podcast Growth

Provides platform-specific optimization, cross-promotion, and listener acquisition strategies for podcast growth. Used by distribution managers to maximize discoverability and audience engagement.
SEO
+0%1.2K419