The Best Design & Creative Skills for Claude Code (Diagrams, Images, Infographics)
The creative category has the widest quality spread in the ecosystem — for every skill that produces genuinely professional output, three produce generic AI slop. Here are the seven that cleared the bar.
Design skills live or die on output quality, which makes them the hardest category to evaluate — and the easiest to fake. A prompt-wrapper can call itself an "infographic generator." Whether the infographics are actually usable is a different question.
The six-dimension rubric treats content quality and engineering rigor equally, which is what separates this list from a "cool tools" roundup.
1. website-to-video — 9.5/10
Turns any website into a polished demo or explainer video — scrolling shots, feature callouts, smooth transitions.
Why: The most immediately practical skill in the batch. Founders and marketers need product videos constantly; this generates them from the live site with no footage required. The scene sequencing logic is genuinely smart about pacing.
The one thing: Output quality tracks the design quality of the source site. A rough site gets a rough video.
2. baoyu-diagram — 9.4/10
Professional dark-themed SVG diagrams of any type — architecture, flowcharts, sequence diagrams.
Why: SVG output means the result scales, edits, and embeds cleanly anywhere. The generated diagrams have a consistent professional identity instead of looking like each one came from a different tool.
The one thing: Dark theme is the house style. If you need light-theme diagrams for print, check compatibility before committing.
3. baoyu-image-gen — 9.3/10
AI image generation with OpenAI, Azure, Google, and OpenRouter backends.
Why: Backend abstraction done right — one interface, multiple providers, clean fallback. The engineering is what separates it from the hundred "DALL-E wrapper" skills.
The one thing: You need at least one API key configured. No key, no images — this isn't a limitation, it's honest design.
4. baoyu-infographic — 9.3/10
Professional infographics with 21 layout types and 22 visual styles.
Why: The layout engine is real: it picks layouts based on your content structure (comparison, timeline, process, hierarchy) instead of dumping everything into one generic template. Data-driven infographics come out looking designed, not decorated.
The one thing: It works best with structured input. Give it your key points as data, not as prose paragraphs.
5. visual-style — 9.3/10
Create, extract, and apply portable visual design systems via visual-style files.
Why: The most interesting idea in the category: a portable format for design systems that lets your agent apply a consistent visual identity across anything it makes — diagrams, documents, slides. Consistency is the hardest thing in AI-generated design, and this solves it structurally.
The one thing: You need to invest in creating (or extracting) your style file first. The payoff is proportional to how much you generate.
6. baoyu-cover-image — 9.2/10
Article cover images along 5 dimensions — type, palette, rendering style.
Why: Cover images are a repetitive task with real stakes (the cover is the click decision). Dimension-based generation gives you controllable variety instead of slot-machine output.
The one thing: It's tuned for article/blog covers. Social media sizes need manual adjustment.
7. baoyu-comic — 8.9/10
Knowledge comics in multiple art styles and tones.
Why: The rare creative skill that's also pedagogically serious — it thinks about panel pacing and visual metaphor, not just "draw a comic about X."
The one thing: Comics amplify whatever your content actually is. Weak explanations become weak comics, visibly.
The pattern
The winning skills all have a structure layer between the prompt and the output: a layout engine, a style system, a sequencing logic. The losers pass your prompt straight to a generation API and shrug at the result. Structure is what makes creative output professional.
Full scorecards are on the design & creative collection page.