Annotator (#1861)

<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Adds an in-app screenshot annotator to the Preview panel for Pro users
so you can capture the current app view, draw or add text, and submit an
annotated image to chat.

- **New Features**
- Pen button in PreviewIframe to toggle annotator; captures a screenshot
via worker messaging and displays it in a Konva canvas.
- Tools: select, freehand draw, and draggable text; supports undo/redo,
delete, and resizing with Transformer. Canvas scales to the container.
Includes a color picker.
- Submit exports a PNG and attaches it to the chat via useAttachments;
prefills the chat input; annotator auto-closes after submit.
  - Pro-only: non-Pro users see an upsell screen.
- State atoms added: annotatorModeAtom, screenshotDataUrlAtom,
attachmentsAtom; PreviewIframe now handles dyad-screenshot-response
messages.

- **Dependencies**
  - Added konva, react-konva, perfect-freehand, and html-to-image.
- Proxy now injects html-to-image and the new dyad-screenshot-client.js
for screenshot capture.

<sup>Written for commit 580aca271c5993a0dc7426e36e34393e073bd67b.
Summary will update automatically on new commits.</sup>

<!-- End of auto-generated description by cubic. -->

forge.config.ts

@@ -32,6 +32,9 @@ const ignore = (file: string) => {
   if (file.startsWith("/node_modules/stacktrace-js/dist")) {
     return false;
   }
+  if (file.startsWith("/node_modules/html-to-image")) {
+    return false;
+  }
   if (file.startsWith("/node_modules/better-sqlite3")) {
     return false;
   }