Add a dockable in-app browser panel (like Codex's embedded browser)

Feature request: Dockable in-app browser panel

Currently, Claude in Chrome controls a separate Chrome window. OpenAI's Codex provides an embedded browser panel docked inside the app UI (e.g. on the right side), so the user can see the browser and the agent's actions side-by-side without window juggling.

Request: add an optional dockable/embedded browser view inside Claude Code, so browser automation (navigation, clicks, screenshots) renders in a panel within the app rather than only in a standalone Chrome window.

Why: keeps the agent's browsing visible alongside the conversation, matching the Codex workflow many users expect.