drawio-skill

You can generate flowcharts from text by providing a textual description to the diagram generation process, which creates a JSON render specification and outputs a visual diagram. This skill uses draw.io to render the final architecture visualization from your prompt.

Yes, draw.io works with Claude's vision model to perform visual self-checks on generated diagrams. Claude's vision capabilities evaluate the visual correctness of the diagram generation output and refine the architecture visualization automatically.

The best way to export draw.io diagrams to multiple formats is by using the draw.io desktop app CLI. You can export your generated diagrams to PNG, SVG, PDF, and JPG formats, with an option to embed XML for later editing.

Yes, you need the draw.io desktop app CLI installed to create architecture visualizations with this skill. The diagram generation process requires the CLI to render JSON specifications into visual outputs and export them successfully.

Yes, you can embed XML when exporting diagrams to formats like PNG and SVG. This optional XML embedding ensures the exported file remains editable later in draw.io while retaining the visual correctness checked by Claude's vision.

Visual self-checking works by passing the generated diagram to Claude's vision model to evaluate its visual correctness against the original text description. This mechanism ensures the diagram generation output accurately reflects the requested architecture visualization.