Fix auto-continue bug: don't return early when tools executed but final_output not called

The bug was in the chunk.finished block inside stream_completion_with_tools.
When no tool was executed in the CURRENT iteration (!tool_executed), the code
would return early without checking if tools were executed in PREVIOUS iterations
(any_tool_executed) and final_output was never called.

This caused the agent to terminate prematurely after executing tools like
todo_read when the LLM responded with text instead of calling final_output.

The fix adds a check: if any_tool_executed && !final_output_called, we break
to let the outer loop's auto-continue logic prompt the LLM to continue.

Also fixed missing debug! import in g3-console/src/main.rs.

crates/g3-core/src/lib.rs

@@ -4884,6 +4884,17 @@ impl<W: UiWriter> Agent<W> {
                                     ));
                                 }
 
+                                // If tools were executed in previous iterations but final_output wasn't called,
+                                // break to let the outer loop's auto-continue logic handle it
+                                if any_tool_executed && !final_output_called {
+                                    debug!("Tools were executed but final_output not called - breaking to auto-continue");
+                                    // Add the text response to context before breaking
+                                    if has_text_response && !current_response.trim().is_empty() {
+                                        full_response = current_response.clone();
+                                    }
+                                    break;
+                                }
+
                                 // Set full_response to current_response (don't append)
                                 // current_response already contains everything that was displayed
                                 // Don't set full_response here - it would duplicate the output