bundled-skills/article-illustrations/SKILL.md
Generate hand-drawn 16:9 article illustrations with the Grav character IP, sparse annotations, and absurd but clear visual metaphors.
npx skillsauth add FrancoStino/opencode-skills-antigravity article-illustrationsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Generate 16:9 landscape hand-drawn illustrations for articles, blog posts, and technical content. Each illustration captures one cognitive anchor point from an article and turns it into a clean, absurd, memorable whiteboard-sketch explanation.
The skill uses a recurring character IP called Grav: a small, round, always-floating figure with dot eyes and a thin antenna. Grav participates in the core action of every illustration — never just decoration.
Repository: vipin-si/article-illustrations
Read the article and identify cognitive anchor points — core judgments, turning points, input/output loops, before/after contrasts, and common pitfalls. Don't distribute illustrations evenly; prioritize moments that benefit from visual explanation.
For each illustration, define:
Use the generate_image tool with the built-in prompt template. Each image follows strict style rules:
Verify each image against the QA checklist: correct format, Grav present and active, original metaphor, clean composition, sparse annotations, correct color usage.
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.
<paste article>
Generate one 16:9 illustration for this concept:
"Trust isn't declared — it's built one piece of evidence at a time."
Grav must perform the core action. Maximum 5 annotation labels.
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.
| Element | Rule | |:--------|:-----| | Background | Pure white — no cream, texture, gradients, or shadows | | Line art | Black, hand-drawn, slightly wobbly, not mechanical | | Whitespace | Main subject 40–60% of canvas, 35%+ empty space | | Annotations | Handwritten English, 2–5 words each, max 5–8 per image | | Color: Black | Main line art, characters, structures, objects | | Color: Red | Key highlights, problems, warnings, results | | Color: Orange | Main flow, paths, arrows, direction | | Color: Blue | Supplementary notes, feedback, system state | | Prohibited | Green, purple, yellow, pink, gradients, drop shadows, 3D, realistic UI |
Problem: Illustration looks like a PPT slide Solution: Remove 30% of elements, increase whitespace, make it weirder
Problem: Grav is just standing next to the action Solution: Redesign so Grav IS the mechanism — becomes the funnel, dangles from the lever, is suspended inside the machine
Problem: Same metaphor as a previous illustration Solution: Replace the physical object entirely — same concept, different analogy
tools
Authorized security assessment of LLM applications and AI agents: prompt injection, tool abuse, RAG exposure, memory poisoning, system-prompt extraction, and agent-compliance engineering per OWASP LLM/ASI Top 10.
development
Builds two parameterized UI modes—流光溢彩白 (iridescent white) and 五彩斑斓黑 (colorful black)—with OKLCH, WebGL/CSS fallback, vision gating, screenshot QA, and total/per-color intensity reports. Use when a UI request names either mode or needs measured color parameters.
tools
Delegate coding tasks to the Kimi Code CLI (`kimi`) only when the user explicitly requests it, while the orchestrator retains review and landing responsibility.
development
Front-end JavaScript reverse engineering: locate signature chains, analyze encrypted request parameters, sample runtime behavior, and reproduce logic locally in Node for evidence-based output.