Wednesday afternoon, a tester dropped an issue on me: a certain page threw a 500 every time you clicked Save. It came with a screenshot and reproduction steps.
Normally this would go into the queue. I had other things on my plate, so the earliest I could get to it was that evening. But this time I replied under the issue with one line:
@claude Take a look at this 500, locate it using the repro steps, and if it's fixable, just fix it and open a PR
Then I went back to my own work. Nine minutes later my phone buzzed with a GitHub notification: a PR had been opened. Which file changed, why it was changed that way, and the results of the local tests were all clearly laid out. I skimmed the diff, and it was cleaner than what I’d have written myself.



