What problem does it solve? Creating architecture, workflow, or concept diagrams by hand is slow, and generic AI-generated diagrams tend to be uniform boxes-and-arrows that display labels without communicating structure. This Skill produces Excalidraw JSON files whose visual layout mirrors the concept being explained, in a crisp draw.io / diagrams.net style rather than a hand-drawn look. ## Core Features & Use Cases - Concept-to-pattern mapping: Matches each concept to a visual pattern (fan-out, convergence, timeline, tree, cycle) so the diagram's structure argues the idea, not just labels it. - Evidence artifacts for technical diagrams: Embeds real code snippets, JSON payloads, and actual event or API names researched from specifications so diagrams teach concrete details. - Section-by-section generation for large diagrams: Builds comprehensive diagrams incrementally to avoid output token truncation, with cross-section arrow bindings and a geometry-based self-review checklist (PNG rendering is disabled; the .excalidraw JSON is the deliverable). - Brand-customizable palette: All colors, fonts, and stroke defaults live in references/color-palette.md as a single source of truth you can swap for your own brand. - Use Case: Ask for a diagram of how a streaming protocol delivers events from an agent backend to a frontend UI, and receive a .excalidraw file with real event names, a timeline of the event sequence, and grouped sections for each system component. ## Quick Start Ask your coding agent to create an Excalidraw diagram showing how your system's components communicate, and it will generate the .excalidraw JSON file for you.
