Everything I could decide myself is decided. What's left needs
you. Each one has a recommendation from Fable — if you agree, pick it and
press Done. "I'm not sure" is a real answer and becomes work for me.
You don't need to open the session — pressing Done sends your answer back and it picks up.
1What counts as done for this session — the EXPIRED Slack flood fix plus moving that loop to Temporal? You can skip this page entirely by typing drive "<what done means>" in your own session; that writes it straight to me. It is /drive, not /goal.
In plain terms. This session is fixing the alert spam you saw, and also getting the watchdog behind it onto Temporal so it stops needing a babysitter. It has since found that the Temporal loop already exists and is simply switched off, so that half may be much smaller than expected. I need to know when to call the whole thing finished: only after we actually watch the alerts channel stay quiet and see the loop running, or sooner - the moment the spam stops, with the loop tracked as its own piece of work.
Fable recommends: Both halves, checked live: the alerts channel stays quiet through a full cycle and the loop is running on Temporal with Temporal owning its timer (no cron, no bash poll). You asked for both in one breath and the session took both on, so narrowing it to close sooner would be me shrinking your scope. And for an alerting loop the honest proof is the channel going quiet, not a merged diff. The flood was firing constantly, so one quiet cycle is hours, not days. Tried first, unsuccessfully: receipt f6efc1992 — RESOLVED, held under reversal, high confidence both ways. It found the flood existed BECAUSE an earlier fix trusted the wrong proxy while its unit test passed, which is the exact failure only live watching catches, so closing on merge alone would repeat it. It flagged the cost if this is wrong: a verified predicate fix sits unclosed behind a Temporal cutover if that half stalls. No channel or production risk either way.