webcomic

name: webcomic description: > Generate a webcomic page from a feature description or the current git branch diff. Creates 4-6 illustrated panels with narrative flow, generates each via Gemini, and combines into a single comic page PNG. Triggers on "webcomic", "make a comic", "comic about", or /webcomic. user_invocable: true

Webcomic Generator

Generate a complete webcomic page: write the narrative, generate panel art with Gemini, and assemble into a single comic page image.

Configuration

STYLE: funny
PANELS: 4
MODEL: pro

Workflow

Step 1: Determine Topic

Parse $ARGUMENTS for:

• Style keyword (optional): funny | dramatic | technical | absurd | noir — defaults to funny
• Panel count (optional): a number 4-6 — defaults to 4
• Topic: everything else in the arguments

Style and panel count can appear in any order before the topic. Examples:

• /webcomic the login flow → funny, 4 panels, topic = "the login flow"
• /webcomic dramatic 6 our deploy pipeline → dramatic, 6 panels, topic = "our deploy pipeline"
• /webcomic noir the auth system → noir, 4 panels, topic = "the auth system"

If no arguments are provided, auto-detect the topic from the current branch:

git diff main...HEAD --stat
git diff main...HEAD

Summarize what changed and use that as the comic topic. Use the branch name for additional context.

Step 2: Design Panels

Create a narrative arc with setup → development → climax → resolution. Write a detailed image prompt for each panel.

Critical prompt rules:

• Every prompt MUST include: "No text, no speech bubbles, no words, no letters, no captions in the image."
• Describe the scene visually — characters, setting, action, mood, lighting
• Maintain visual consistency across panels (same characters, same art style)
• Include the style tone in prompts (e.g., "cartoon style, bright colors, exaggerated expressions" for funny)

Style guidance:

Write a short caption (1 sentence) for each panel that tells the story. Captions are overlaid by combine.py, NOT baked into the image.

Present the panel plan to the user and ask for approval before generating.

Step 3: Create Output Directory

mkdir -p ~/Desktop/webcomic-{slug}/

Where {slug} is a kebab-case version of the topic (e.g., webcomic-login-flow).

Step 4: Generate Panel Images

Generate all panels using image.py. Run them in parallel (separate Bash calls) for speed.

For each panel:

uv run ~/.claude/skills/webcomic/scripts/image.py \
  --prompt "{panel_prompt}" \
  --output ~/Desktop/webcomic-{slug}/panel-{N}.png \
  --aspect landscape \
  --model pro \
  --size 1K

If a panel fails, retry it once. If it fails again, skip it and note the failure.

Step 5: Combine into Comic Page

uv run ~/.claude/skills/webcomic/scripts/combine.py \
  --title "{comic_title}" \
  --panels ~/Desktop/webcomic-{slug}/panel-1.png ~/Desktop/webcomic-{slug}/panel-2.png ... \
  --captions "Caption 1" "Caption 2" ... \
  --output ~/Desktop/webcomic-{slug}/comic-page.png \
  --style {style}

Step 6: Show Result

open ~/Desktop/webcomic-{slug}/comic-page.png

Report the output path and a summary of what was created.