I threw the bug-fixing task into the Slack group, and half an hour later Claude tossed back a PR.
Last Wednesday at ten in the morning, customer support dropped a screenshot into the group chat: a user clicked “Export Report,” the page spun for three rotations, and finally popped up the message Internal Server Error. Two more messages followed right below it, one with the user ID and one with reproduction steps added by the operations team.
In the old days, I knew this workflow all too well: copy the chat history, open Jira and create a ticket, switch to the terminal, dig through logs, ask support for an account, then go back to the group to sync progress. The bug hadn’t even started getting fixed, and the context had already been carried back and forth three times.
This time I didn’t switch to anything. I just replied in Slack with one line:
@Claude Please look into this export error. The relevant endpoint is
reports/export, and the permission check was just changed yesterday. Reproduce it first, then fix it, add a regression test, and finally open a PR.
