Automatically picks the right diagram format based on what you're trying to show. Feed it a topic, file, or recent conversation and it classifies the content shape (graph, sequence, state machine, comparison, etc.) then renders it as Mermaid, ASCII art, or a table depending on where you're viewing it. The decision table for content detection is smart: architecture gets flowcharts, API calls get sequence diagrams, tradeoffs get markdown tables. Works best when you know something needs a visual but don't want to think about whether it should be a state diagram or an ER model. Falls back gracefully when Mermaid isn't available, and you can override with a specific format if the auto-detection misses.
npx skills add https://github.com/camacho/ai-skills --skill visualize