Design Mode is open in the Agents Window browser and an agent is midway through a spacing change when a second problem becomes obvious elsewhere on the page. What does Cursor say about sending that second edit straight away?

The principle — Design Mode is built for the rhythm of interface review, where one fix leads straight to noticing the next. Rather than making you wait for each edit to land before you can describe another, it lets you dispatch changes as you spot them.

Why the key is correct — Cursor describes the flow directly: point at one element, describe the change, move to another part of the page, and send another edit before the first one finishes. Each visual prompt becomes its own subagent, so several are in flight together, and as each finishes the running app hot reloads and the change appears in front of you. The same design shows up elsewhere in the mode, where the microphone stays available while agents are running so the next instruction can be queued without a pause.

Why the others are wrong — Expecting a second instruction to be folded into the running edit treats the surface as a single conversation with a single worker, which is the ordinary chat model rather than this one. Expecting it to supersede and discard the first change adds a destructive behaviour the documentation does not describe, and that fear is precisely what makes people serialise. Treating the second edit as merely queued accepts the dispatch but keeps the serial timing, leaving you idle through the very stretch the mode was designed to fill.

Remember this — In Design Mode, keep moving. Send the next edit while the last one is still running.

Sources — Cursor's Design Mode documentation.