add shikane-watchdog user service to auto-recover shikane deadlocks

README.md

@@ -11,3 +11,25 @@ stow .
 ```
 If stow is not installed use ./install.sh from .dotfiles repo
 
+# User services
+
+## shikane-watchdog
+
+`.config/systemd/user/shikane-watchdog.timer` (every 2 min, user unit)
+runs `.config/sway/scripts/shikane-watchdog.sh`, which restarts the
+`shikane` user service if its state machine is stuck.
+
+Why: shikane 1.1.1 deadlocks when a profile application is discarded
+mid-way (e.g. by a failing DP atomic commit on the external display).
+The daemon then keeps logging
+`[WARN shikane::variant] Received invalid input AtomicChangeDone at state Discarded`
+and `shikanectl reload` no longer recovers it. The watchdog detects
+that signature in the journal and runs `systemctl --user restart shikane`
+(re-applies the active profile in ~1s). It never restarts a healthy
+or legitimately transitioning daemon.
+
+```console
+systemctl --user status shikane-watchdog.timer      # check it's active
+systemctl --user disable --now shikane-watchdog.timer  # disable
+```
+