An Antigravity skill that turns the key judgments, processes, and metaphors in your articles into memorable 16:9 hand-drawn illustrations — featuring Grav, a floating character who earnestly participates in absurd system diagrams.

Article Illustrations is an Antigravity skill that guides an AI agent to design and generate inline article illustrations.

It's not a generic illustration prompt, and it's not a PPT infographic template. The core goal is: understand the cognitive anchor points in an article, then turn one judgment, process, structure, state, or metaphor into a memorable 16:9 hand-drawn explanation sketch.

The default visual IP is Grav: a small, round, always-slightly-floating figure with dot eyes and a single thin antenna. Grav hovers just above surfaces — never quite touching the ground — like a quiet systems engineer who defies gravity while earnestly doing mundane work. Grav isn't a mascot or sticker — Grav is an absurd worker participating in whatever system the illustration depicts.

In one line: Make the AI not just "add an image" but actually draw out a key cognitive action from the article.

Great fit:

• Writers who need inline illustrations for technical articles and blog posts
• Knowledge workers creating methodology content, AI workflow content, or system design content
• People who want to turn abstract judgments into concrete visual metaphors
• People who want a lighter, weirder, more distinctive illustration style than PPT infographics
• Anyone using Antigravity for content production who wants a consistent visual language

Not a fit:

• Commercial illustration, brand KVs, or polished flat illustration
• Traditional PPT infographics, complex architecture diagrams, or flowcharts
• Children's cartoons, cute IP, or emoji/sticker styles
• Cramming lots of text, long explanations, or full course pages into one image
• Strictly editable vector source files

Default output:

• 16:9 landscape inline article illustrations
• A shot list of 4–8 illustrations per article
• For each: theme, core meaning, structure type, Grav's action, and annotation suggestions
• Final PNG images

Does not produce:

• PPTX / PDF / Keynote
• SVG / HTML / Canvas editable graphics
• Commercial posters or cover art
• Dense text-heavy infographics

This skill uses a distinctive hand-drawn style:

• Pure white background — no textures, gradients, or shadows
• Black hand-drawn line art — thin lines, slight wobble, not mechanical
• Generous whitespace — main subject occupies ~40%–60% of canvas
• Sparse red/orange/blue handwritten English annotations
• One image = one core action, structure, state, or metaphor
• Grav must participate in the core action, never just decorate
• Grav always floats — never touches surfaces
• Absurd, creative, clean — but never childish or cute

These illustrations demonstrate the Grav style across different universal concepts. Each one was generated using this skill.

Too many inputs, one brain. Grav holds a tiny bucket as pipes pour envelopes, bells, and chat bubbles from every direction.

Trust isn't declared — it's built one brick at a time across a chasm. Grav floats over the gap, placing bricks labeled "proof" and "track record."

A cloud of lightbulbs funnels into a press machine. Grav pulls the lever. One diamond comes out; crumpled rejects fall below.

Build → Measure → Learn → Repeat. Grav carries an insight nugget from the "learn" stage back to "build." Cycle 4.

These are style examples, not composition templates. The skill generates original metaphors for each article.

Clone the repo:

git clone https://github.com/vssinghh/article-illustrations.git
cd article-illustrations

Copy the skill to your Antigravity skills directory:

mkdir -p ~/.gemini/config/plugins/custom-skills/skills/article-illustrations
cp SKILL.md ~/.gemini/config/plugins/custom-skills/skills/article-illustrations/

Or symlink it:

ln -s "$(pwd)" ~/.gemini/config/plugins/custom-skills/skills/article-illustrations

Analyze this article and create a shot list of 5 illustrations.
Don't generate images yet — just plan which cognitive anchor points
deserve illustrations and what each image should convey.

<paste article>

Generate 4 Grav-style illustrations for this article.
Requirements: 16:9 landscape, pure white background, black hand-drawn
line art, sparse red/orange/blue English annotations.
Each image should capture one core structure.

<paste article>

Generate one 16:9 illustration for this concept:

"Trust isn't declared — it's built one piece of evidence at a time."

The scene should be absurd but clean. Grav must perform the core action.
Maximum 5 annotation labels, keep them short.

This illustration is on the right track, but Grav feels like decoration.
Keep the core meaning but regenerate: make Grav the one actually
driving the structure. Make the scene weirder but still clean and minimal.

Create an illustration strategy for this long article.
Don't distribute illustrations evenly — only pick cognitive anchor points:
core judgments, input/output loops, before/after contrasts, common pitfalls.
Plan 6-8 images, output the shot list first, don't generate yet.

<paste article>

This skill was inspired by ian-xiaohei-illustrations, a Codex skill for generating Chinese article illustrations with a detailed multi-file reference system (style DNA, character IP sheets, composition patterns, prompt templates, and QA checklists). We simplified that approach into a single self-contained skill file, adapted it for English articles, introduced an original character IP (Grav), and made it work natively with the Antigravity platform.

MIT — see LICENSE.