fix(tui): preserve watchdog behavior on refreshed base

Resolve the remaining TUI-side merge drift after restacking the watchdog and subagent behavior branch onto the refreshed TUI foundation branch.

Co-authored-by: Codex <noreply@openai.com>

codex-rs/tui/src/chatwidget.rs

@@ -3734,6 +3734,9 @@ impl ChatWidget {
                             prompt: prompt.unwrap_or_default(),
                             model: String::new(),
                             reasoning_effort: ReasoningEffortConfig::Medium,
+                            // Thread history items do not carry spawn_mode yet, so the
+                            // replay path must choose an explicit fallback for reconstructed
+                            // spawn rows. Plain spawn is the least surprising default.
                             spawn_mode: AgentSpawnMode::Spawn,
                             status: first_receiver
                                 .as_ref()