What problem does it solve? Working with open tldraw Desktop documents programmatically is hard because the canvas state lives inside a running app. This Skill exposes the app's local HTTP server so an agent can list documents, read shapes and bindings, execute JavaScript against a live editor, capture screenshots, and install durable document scripts without manual clicking. ## Core Features & Use Cases - Canvas inspection and editing: Discover open documents, read shape records and bindings, and mutate the canvas via the /api/search and /api/doc/:id/exec endpoints. - Durable behavior scripts: Write script/main.js and script/config.js through the script workspace to add clickable UI, animations, custom shapes, bindings, and overlays that persist with the document. - Diagram quality checks: Create properly bound arrows with helpers.createArrowBetweenShapes and run helpers.getLints() before reporting a diagram complete. - Use Case: Ask the agent to build a project plan diagram in a new tldraw document — it creates the file via /api/docs/create, draws labeled boxes with real arrow bindings, lints the result, and saves the document. ## Quick Start Ask the agent to list my open tldraw documents and add a labeled rectangle to the focused canvas.
