feat: add ImagePasteAddon for clipboard image handling

[brianegan]

Summary

• Adds ImagePasteAddon, a new addon following the ITerminalAddon pattern (same as FitAddon), that detects image data in clipboard paste events and emits them as base64-encoded payloads via an onImagePaste event
• Updates InputHandler.handlePaste to only claim paste events that contain text, allowing image-only pastes to bubble through to addons
• Keeps the core Terminal API strictly xterm.js-conformant — no custom events on the Terminal class

Usage

import { Terminal, ImagePasteAddon } from 'ghostty-web';

const term = new Terminal();
const imagePasteAddon = new ImagePasteAddon();
term.loadAddon(imagePasteAddon);

imagePasteAddon.onImagePaste((data) => {
  console.log(data.name);       // e.g. "clipboard_1234567890.png"
  console.log(data.dataBase64); // base64-encoded image data
});

Design decisions

• Addon, not core API: onImagePaste doesn't exist in xterm.js. Implementing it as an addon keeps the drop-in replacement promise intact — code written against the core API remains portable between xterm.js and ghostty-web.
• Event bubbling: InputHandler no longer calls preventDefault()/stopPropagation() at the top of paste handling. It only claims the event when there is text to process, letting image-only pastes reach addon listeners.
• Exported types: ImagePasteAddon and IImagePasteData are exported from the package entry point.

Test plan

• All 341 existing tests pass
• New image-paste.test.ts covers activation, disposal, lifecycle, event subscription, image paste firing, non-image paste ignoring, and post-dispose cleanup
• Manual test: paste an image from clipboard with the addon loaded and verify onImagePaste fires
• Manual test: paste text with the addon loaded and verify normal onData behavior is unaffected

🤖 Generated with Claude Code

[brianegan]

Heya, really cool project! I was building a VSCode extension with ghostty-web and ran into the issue where I needed to paste images. I created this plugin with Claude Opus to handle such a case.

Not sure if you're accepting submissions coauthored by AI, but if so, I've found this helpful and wanted to contribute.