fix(gateway): reap orphans on stop_profile_gateway (#75936)

stop_profile_gateway() only killed the single PID recorded in the pid
file. On repeated restarts, the new process overwrites the pid file
before the old one exits, making older gateway instances invisible
to subsequent stops. Each restart stacked another orphan — observed
as N threads and N replies per Discord message.

After killing the recorded PID, also call _reap_unsupervised_gateway_orphans()
to sweep any remaining gateway processes for this profile. The reap function
already handles the no-systemd guard and SIGTERM+SIGKILL escalation.

Fixes #75936

hermes_cli/gateway.py

@@ -1583,6 +1583,12 @@ def stop_profile_gateway() -> bool:
     a live orphan still holds the webhook port. In that case fall back to the
     orphan-aware process scan so the replacement reaps the prior instance
     instead of stacking a duplicate on the same port (#51325).
+
+    Even when the pid file is valid and points to the current gateway, older
+    orphans may linger from prior restarts that overwrote the pid file before
+    the old process exited. After killing the recorded PID, also sweep for
+    any remaining orphans so each restart produces at most one live gateway
+    (#75936).
     """
     try:
         from gateway.status import get_running_pid, remove_pid_file
@@ -1620,6 +1626,14 @@ def stop_profile_gateway() -> bool:
 
     if get_running_pid() is None:
         remove_pid_file()
+
+    # Also reap any orphans from prior restarts whose PIDs were overwritten
+    # in the pid file before they exited (#75936).
+    try:
+        _reap_unsupervised_gateway_orphans()
+    except Exception:
+        pass
+
     return True