Everything I could decide myself is decided. What's left needs
you. Each one carries the recommendation of the session that raised it — if you
agree, pick it and press Done. A model's opinion, where there is one, appears on the card
as a receipt line; cards with no receipt were never put to a model. "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.
1The mini has 93 idle background sessions whose kickoff prompt is the claude agents --json roster (the roster-as-prompt spawn loop), started 03:38–04:31 CT at ~1/min. They hold ~28 GB; swap is 20.4 of 21.5 GB used and load averaged 200+. Today that stalled three Smith jobs (sensor-expiry-audit, graphify-refresh, worker eviction → self-restart at 08:45 CT). Kill the idle roster-prompted sessions now?
In plain terms. Should I shut down the 93 idle robot sessions that are clogging the Mac mini right now?
The session raising this recommends: Kill them now They are idle, roster-prompted junk from a known loop (memory note reference_roster_as_prompt_spawn_loop, since 08-07); each holds ~300 MB; nothing is driving them; the spawner is the claude daemon dispatch, outside Smith's edit scope, so killing is relief not a fix. Tried first, unsuccessfully: Measured: 152 roster peers, 93 named pr-deps*/pr-deps-three*, 3 non-idle; vm.swapusage 20456/21504 MB; load 95/213/286; three 120–180s subprocess timeouts and a Temporal eviction wedge in the worker log 07:41–08:04 CT. Did not kill anything: they are other sessions, not mine.