Notify on run failure
D-181 slice 5b § 12 — the background-failure notify handler dish. Subscribes to the D-179 § 8.4 run-outcome bus family (`run.<recipe_id>.<dish_id>.failed`) and dispatches one D-158 notification per failed run so a failure nobody was watching (a scheduled / reactive / housekeeping run) still surfaces. The bridge_mode gate is the delivery layer's, not the recipe's: a `notification-send` `notify` fans out only to bridges that hold notification mode (`routeBridgeNotify`), so enabling/disabling per-bridge push is Settings → Notifications, not this recipe. Self-failures don't storm: the dispatcher's `origin_trigger_id` guard suppresses a trigger re-firing the run it dispatched, and `skip_when` belts-and-suspenders the no-origin (manually-run) edge by skipping when the failed run IS this recipe.
How it works
Inspect the data fetches, transforms, gates, and output this recipe runs.
Settings
Configurable at install. Defaults shown — change them anytime in Recued.
channels
setting
=
in_app
Trust & control
What installing this recipe would let it do. Recued grants these permissions at install — review them there before approving.