fix: prevent fallback PID from killing external server in managed mode

When REMOTE_MANAGED is 'managed' but REMOTE_PID is empty (e.g. PID file
deleted or corrupted), the previous code fell back to DEFAULT_RUNTIME_PID
via ${REMOTE_PID:-$DEFAULT_RUNTIME_PID}. This could kill an externally
started server, causing wait_ready to fail and triggering an unnecessary
full relaunch cycle.

Fix: use REMOTE_PID directly without fallback so we only kill the process
we actually launched. If REMOTE_PID is empty, the -n guard skips the kill.

Author: cursoragent

packages/ssh/src/tunnel.ts

@@ -546,7 +546,7 @@ if [ -n "$DEFAULT_REMOTE_PORT" ]; then
   REMOTE_PORT="$DEFAULT_REMOTE_PORT"
   if wait_ready "@@T3_REUSE_READY_TIMEOUT_MS@@"; then
     if [ "$REMOTE_MANAGED" = "managed" ]; then
-      PID_TO_STOP="\${REMOTE_PID:-$DEFAULT_RUNTIME_PID}"
+      PID_TO_STOP="$REMOTE_PID"
       if [ -n "$PID_TO_STOP" ] && kill -0 "$PID_TO_STOP" 2>/dev/null; then
         kill "$PID_TO_STOP" 2>/dev/null || true
         wait_for_pid_exit "$PID_TO_STOP"