Fix TUI status visibility around streamed output (#33105)

## What changed

- Hide the working-status indicator when finalized assistant output is committed to the transcript, including single-line final answers.
- Restore the indicator when image generation begins during a running task after its streamed preamble has been flushed.

## Testing

- Add snapshot coverage for a single-line final answer and for image generation following a single-line preamble.

GitOrigin-RevId: e294a11cda3f0ce46f2bbed1be2a1380c9c73085

codex-rs/tui/src/chatwidget/snapshots/codex_tui__chatwidget__tests__image_generation_begin_restores_working_status.snap

@@ -0,0 +1,11 @@
+---
+source: tui/src/chatwidget/tests/exec_flow.rs
+expression: normalized_backend_snapshot(terminal.backend())
+---
+"                                                                                "
+"• Working (0s • esc to interrupt)                                               "
+"                                                                                "
+"                                                                                "
+"› Ask Codex to do anything                                                      "
+"                                                                                "
+"  gpt-5.6-sol default · /tmp/project                                            "

codex-rs/tui/src/chatwidget/snapshots/codex_tui__chatwidget__tests__single_line_final_answer_hides_working_status.snap

@@ -0,0 +1,14 @@
+---
+source: tui/src/chatwidget/tests/status_and_layout.rs
+expression: normalize_snapshot_paths(terminal.backend().vt100().screen().contents())
+---
+
+› count to 1
+
+
+• 1
+
+
+› Ask Codex to do anything
+
+  gpt-5 default · /tmp/project

codex-rs/tui/src/chatwidget/streaming.rs

@@ -26,6 +26,11 @@ impl ChatWidget {
             };
             self.clear_active_stream_tail();
             let (cell, source) = controller.finalize();
+            // Match newline-committed streaming behavior: once assistant output is ready to be
+            // committed into history, hide the inline status row so transcript content replaces it.
+            if cell.is_some() {
+                self.bottom_pane.hide_status_indicator();
+            }
             let deferred_history_cell =
                 if scrollback_reflow == crate::app_event::ConsolidationScrollbackReflow::Required {
                     cell

codex-rs/tui/src/chatwidget/tests/exec_flow.rs

@@ -877,6 +877,30 @@ async fn view_image_tool_call_preserves_foreign_path() {
     assert_chatwidget_snapshot!("foreign_image_attachment_history_snapshot", combined);
 }
 
+#[tokio::test]
+async fn image_generation_begin_restores_working_status_after_single_line_preamble() {
+    let (mut chat, _rx, _op_rx) = make_chatwidget_manual(/*model_override*/ None).await;
+
+    chat.on_task_started();
+    chat.on_agent_message_delta("Generating an image.".to_string());
+    chat.on_image_generation_begin();
+
+    assert!(chat.bottom_pane.is_task_running());
+    assert!(chat.bottom_pane.status_indicator_visible());
+
+    let width: u16 = 80;
+    let height = chat.desired_height(width);
+    let mut terminal = ratatui::Terminal::new(ratatui::backend::TestBackend::new(width, height))
+        .expect("create terminal");
+    terminal
+        .draw(|frame| chat.render(frame.area(), frame.buffer_mut()))
+        .expect("draw image generation status");
+    assert_chatwidget_snapshot!(
+        "image_generation_begin_restores_working_status",
+        normalized_backend_snapshot(terminal.backend())
+    );
+}
+
 #[tokio::test]
 async fn image_generation_call_adds_history_cell() {
     let (mut chat, mut rx, _op_rx) = make_chatwidget_manual(/*model_override*/ None).await;

codex-rs/tui/src/chatwidget/tests/status_and_layout.rs

@@ -1525,6 +1525,45 @@ async fn streaming_final_answer_keeps_task_running_state() {
     assert!(!chat.bottom_pane.quit_shortcut_hint_visible());
 }
 
+#[tokio::test]
+async fn single_line_final_answer_hides_working_status_snapshot() {
+    let (mut chat, mut rx, _op_rx) = make_chatwidget_manual(Some("gpt-5")).await;
+    chat.thread_id = Some(ThreadId::new());
+
+    complete_user_message(&mut chat, "user-1", "count to 1");
+    chat.on_task_started();
+    complete_assistant_message(
+        &mut chat,
+        "msg-final-single-line",
+        "1",
+        Some(MessagePhase::FinalAnswer),
+    );
+
+    assert!(chat.bottom_pane.is_task_running());
+    assert!(!chat.bottom_pane.status_indicator_visible());
+
+    let width: u16 = 40;
+    let vt_height: u16 = 10;
+    let ui_height = chat.desired_height(width);
+    let viewport = Rect::new(0, vt_height - ui_height - 1, width, ui_height);
+    let backend = VT100Backend::new(width, vt_height);
+    let mut terminal = crate::custom_terminal::Terminal::with_options(backend).expect("terminal");
+    terminal.set_viewport_area(viewport);
+
+    for lines in drain_insert_history(&mut rx) {
+        crate::insert_history::insert_history_lines(&mut terminal, lines)
+            .expect("insert history lines");
+    }
+
+    terminal
+        .draw(|frame| chat.render(frame.area(), frame.buffer_mut()))
+        .expect("draw final answer");
+    assert_chatwidget_snapshot!(
+        "single_line_final_answer_hides_working_status",
+        normalize_snapshot_paths(terminal.backend().vt100().screen().contents())
+    );
+}
+
 #[tokio::test]
 async fn ctrl_c_interrupt_pauses_active_goal_turn() {
     let (mut chat, mut rx, mut op_rx) = make_chatwidget_manual(/*model_override*/ None).await;

codex-rs/tui/src/chatwidget/tool_lifecycle.rs

@@ -25,6 +25,9 @@ impl ChatWidget {
     pub(super) fn on_image_generation_begin(&mut self) {
         self.record_visible_turn_activity();
         self.flush_answer_stream_with_separator();
+        if self.bottom_pane.is_task_running() {
+            self.bottom_pane.ensure_status_indicator();
+        }
     }
 
     pub(super) fn on_image_generation_end(