from draw-io-skill29
Create, edit, lint, and export draw.io diagrams; includes export helpers, SVG linting, and repository QA for clean architecture and icon-based diagrams.
The draw.io Diagram Skill gives an agent the ability to produce and refine native .drawio diagrams, export to PNG/SVG/PDF/JPG, and run repository-grade SVG linting to catch layout issues like border overlap, label collisions, and text overflow. It includes helpers for AWS icons, fixture-based lint coverage, and commands to build and verify documentation.
Use this skill when the user requests architecture diagrams, flowcharts, or visual assets that must be precise and production-ready. Triggers include: "create a draw.io diagram", "export architecture.drawio to svg", "run svg lint on diagram", "find AWS icon ec2", or "refactor login-flow.drawio". It's useful for docs, slide assets, and engineering architecture reviews.
Suited for Claude Code, Codex, and other assistants that can run repository scripts or call CLI tools. The skill is repo-focused and intended for environments where the assistant can read/write .drawio files and run Node/CLI helpers.
This is a comprehensive draw.io diagram skill covering native XML creation, export to PNG/SVG/PDF, and uniquely, SVG linting for diagram quality assurance (overlap detection, text overflow, short terminals, label collisions). The SKILL.md is exceptionally detailed with clear workflow steps, XML rules, and a checklist. Scripts are well-structured with proper error handling and fixture-based verification tests. The Python AWS icon search script expects a service name argument and exits cleanly with usage help when missing. The .mjs scripts were skipped by the runner but appear production-quality from static analysis.
drawio (draw.io desktop CLI — optional, needed only for export commands)Very polished skill. The SVG linting feature is genuinely unique and solves a real problem (draw.io doesn't catch these issues natively). SKILL.md is quite long — could benefit from moving detailed XML/layout rules (section 7) to references/ for better progressive disclosure. The fixture-based verification pattern for the lint checks is excellent engineering practice. Japanese localization support is a nice touch.
