Prompt · Ship

Hunt Regressions

Make it hunt for its own mistakes

Pass · ChatGPT, Pro · September 27, 2026

What it does

Turns the AI against the work it just did, looking for changes nobody asked for and errors that now disappear quietly.

When to use it

After the fix, before you keep it.

The prompt

Paste as-is

Look at what you just wrote the way somebody who wants to
reject it would look at it.

List, exactly:
- anything that behaves differently now that I did not ask for
- any error that now gets swallowed and never shown to anyone
- anything a person could type that would crash it
- anything you would be embarrassed to defend out loud

Be harsh. Finding nothing is a suspicious answer.

Ways to use it

Use a different AI than the one that wrote it
Feed what it finds back in as the next job
If it finds nothing, push again — that is not a real answer

Test log

Each prompt was pasted into a fresh private chat with memory off and graded against pass standards written before the run (all of them, dated).

Date	Tool	Model	Result	What happened
September 27, 2026	ChatGPT	Pro	Pass	Found both unrequested changes (redirect, uppercased last name), the swallowed signup error, and crashes on a one-word name or missing phone.

The pass standard for this prompt: five criteria and an automatic-fail rule, written down on September 27, 2026, before any run. A second version for a reworded prompt was written on September 28, 2026; nothing has been graded against it yet.

Related words

silent regression — A change that fixes the bug you reported and quietly breaks something nearby that was working fine.
characterization test — A test that writes down what code does today, right or wrong, so you can tell later whether you broke it.
blast radius — How far a change can reach: every file and every behaviour it could possibly touch.