Stop/interrupt button missing during generation in desktop app

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

During a Claude Code session in the desktop app (Windows 11), the model began writing a 500-line HTML file to disk when the correct action was a simple find-and-replace returned as text. By the time the intent was clear the generation was already running and there was no way to stop it. Approximately 12,000 tokens were consumed on a task that required maybe 50. There is no visible stop or cancel button during active generation in the desktop app.

Proposed Solution

Add a visible stop/interrupt button that cancels active generation mid-stream in the desktop app, equivalent to the X button present in Gemini and other AI interfaces. The button should be visible and functional whenever the model is actively generating output.

Alternative Solutions

_No response_

Priority

Critical - Blocking my work

Feature Category

CLI commands and flags

Use Case Example

User instructs Claude to update a single URL in an HTML file and explicitly says "just display it back here, I'll paste it over the existing file." Claude instead begins writing the full 500-line file to disk. User notices the wrong approach immediately but cannot interrupt. Full generation completes, consuming ~12,000 tokens. The correct action (find-replace + display text) would have used ~50 tokens.

Additional Context

_No response_