This skill enables any coding agent that reads from a skills directory (like Claude Code or OpenCode) to produce structured Excalidraw diagrams from text prompts. It includes a render pipeline with Playwright so the agent can visually validate and iteratively fix its own outputs before delivering. Diagrams are designed to be semantically visual rather than generic boxes-and-arrows, with evidence artifacts like real code snippets and JSON payloads embedded. Colors and brand styles are centralized for easy customization.