I had a lot of unnoticed inefficiencies in my codebase. Redundant calls, full table scans, N+1 loops quietly firing on every page load. None of them were breaking anything. That's the dangerous part. Everything just worked a little slower than it should have, and every month the database bill was a little higher than it needed to be.

The real problem is that most of these issues are invisible during development. You write a query, it returns the right data, you move on. You don't notice the missing index until you're serving 10,000 users. You don't notice the N+1 until your query count is off by a factor of 100.

So I built a prompt that does the analysis for you.

What the Prompt Actually Does

Instead of manually reviewing every query, you paste your code into the prompt and let the model work through it systematically. It looks for the most common and costly performance problems:

• N+1 query patterns (a new DB call fired for every item in a loop)
• Missing indexes on columns you're filtering or joining on
• Full table scans caused by unindexed or poorly structured queries
• Over-fetching (selecting `SELECT ` when you only need two columns)
• Redundant or duplicate queries hitting the same data twice
• Missing pagination on large dataset queries