
from excalidraw-skill22
Generate hand-drawn Excalidraw-style diagrams (SVG/PNG) from text prompts with layout rules, element format, and a renderer script.
This skill generates Excalidraw-style, hand-drawn diagrams from textual descriptions. It parses user intent into components, relationships, layout direction and outputs diagram JSON that a renderer produces as SVG and PNG. The SKILL.md contains configuration for output directories, element format rules, rendering details, and example workflows.
Use when a user asks for architecture sketches, flowcharts, system designs, deployment diagrams, or any hand-drawn visual. It is invoked for prompts requesting diagrams or via a user command like /excalidraw. Ideal for quickly producing visuals for documentation or design reviews.
Agents capable of invoking Node.js renderers or producing diagram JSON (local assistants, Code/Copilot-like agents, or any agent that can write files and run node).
Excalidraw skill generates hand-drawn style SVG+PNG diagrams from JSON element definitions. The SKILL.md is thorough with clear 5-step workflow, usage examples, and a common mistakes section. The render.mjs script is well-written (~400 lines) with proper text layout, collision avoidance, and font embedding. Scripts couldn't be run because the auditor only supports .py/.sh/.js extensions, not .mjs. No security concerns — purely local rendering with no network calls or credential handling.
node_modules (roughjs, @xmldom/xmldom, @resvg/resvg-js) require npm install)Clean, well-structured skill. The .mjs extension for the renderer means the audit runner couldn't execute it, but static analysis shows it's safe and well-written. The skill fills a useful niche for generating diagrams from text descriptions. Consider whether the auditor should support .mjs extension.