Practice · debugging with evidence
A screenshot and a local route are useful context when the AI tool can access your project. They do not give a remote AI access to your computer. Add what you observed, then compare your report with the example.
Fictional practice scenario
In a neighborhood book-swap app, a signed-in member opens an available book and chooses “Request to borrow.” The button says “Request sent,” but after a refresh the request is missing. The browser console shows POST /api/requests → 401. The form validation changed yesterday.
Loading your saved draft…
Review before you use it
Help me investigate this bug in the project you can actually access. A local URL is only a location hint; it does not grant access to my computer. If you cannot view the route or evidence, ask me to provide what you need. Page or route: [not provided] Steps to reproduce: [not provided] Expected result: [not provided] What actually happened: [not provided] Evidence: [not provided] Recent change: [not provided] First, separate the observed facts from possible causes. Suggest the smallest safe investigation and explain what evidence would confirm or rule out each likely cause. Do not change files, run commands, connect services, or deploy until I approve a proposed fix. After approval, make one focused change, explain what changed, and show how to verify the expected behavior and relevant edge cases.
Your draft stays in this browser and is not sent to an AI service. Use only the fictional scenario here; do not enter passwords, API keys, or private project data. If you copy and paste the prompt into another AI tool, that tool receives it under its own privacy practices.