Use Cursor to spot prototype gaps and draft a build plan

A founder used Cursor to inspect a trip-planning app, surface missing pieces, sketch screens, and then build a feature with parallel agents.

Done with Cursor· Intermediate

What they did
Eric Xiao cloned his trip-planner repository and asked Cursor to explain how it worked. He saved that overview in an AGENTS.md file so later chats would inherit the same context. Next, he had the agent suggest feature ideas and draw ASCII layouts, then split work across parallel agents to review trip history and an itinerary editor before switching to plan mode and worktrees for implementation.
How it went
Cursor surfaced gaps like missing saved trips, no loading state, and no map or follow-up flow. It also produced wireframes for those screens. The post does not show the finished build or any timing details beyond Eric’s claim that he reached a working prototype in about 30 minutes.
Tips
Keep the repo in branches or worktrees and start from a read-only clone so mistakes are easy to roll back. Ask the agent to pause before any code changes, and review the plan before it edits anything.

Try it yourself with Cursor

I’m not a developer. Review [repo or folder] for my [app or prototype] and tell me what is still missing for a usable [feature idea]. Show the missing pieces as a short checklist, then draw simple ASCII wireframes for the main screens. After that, outline a step-by-step build plan and ask for my approval before you change any code. If it helps, use separate worktrees or parallel agents for [part A] and [part B]. Stop and wait before any irreversible action, including deleting files, submitting changes, or anything that would affect users or payments.

Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.

Read the original on amankhan1.substack.com ↗

Get 3 agent uses like this every morning

With the day's top AI news. Free, and one click to unsubscribe.