tl;dr: Creating a hunting team works again, and our API stopped hiding its failures behind a shrug.
Creating a private hunting team was broken. Not slow, not flaky — dead, for eleven days, for everyone. If you tried to start a team and got a vague "something went wrong," that was this. It's fixed, the fix is already live, and no data was lost along the way.
The cause was one line in a database permission rule. Your team record was being created just fine — but the app was then blocked from reading back the row it had literally just written, so the whole operation rolled over and died. Public teams slipped through on a technicality, which is why this looked like an intermittent glitch instead of a hard outage. We restored the missing permission branch, verified the full create-team-and-add-member flow end to end on production for a brand-new account, and left an automated probe in place that fails loudly if this rule ever regresses again. We also swept every other place in the app that writes-then-reads a record: all clear.
The more useful part is why nobody noticed for eleven days. Our mobile API answered every unexpected failure with a single generic message and no error code — nothing to search for, nothing to alert on, and to the app it looked like a dropped connection. Server errors now carry a stable error code distinguishing a permissions denial from a genuine server fault, and we log the underlying database status and the exact route behind it. The message you see stays deliberately vague — raw database text leaks things it shouldn't — but on our side, the same silent failure would now surface in minutes instead of a week and a half.
What this unlocks is boring in the best way: you can create a team, private or public, on web or mobile, and if anything ever does go sideways we'll know before you have to tell us.
Eleven days is eleven days too long. The fix is out, the alarm is wired, and we'd rather tell you about it than hope you didn't notice.