feat(gateway): expose subagent tool calls and thinking to user (fixes #169)

agent/display.py

@@ -199,6 +199,21 @@ def start(self):
     def update_text(self, new_message: str):
         self.message = new_message
 
+    def print_above(self, text: str):
+        """Print a line above the spinner without disrupting animation.
+
+        Clears the current spinner line, prints the text, and lets the
+        next animation tick redraw the spinner on the line below.
+        Thread-safe: uses the captured stdout reference (self._out).
+        Works inside redirect_stdout(devnull) because _write bypasses
+        sys.stdout and writes to the stdout captured at spinner creation.
+        """
+        if not self.running:
+            self._write(f"  {text}", flush=True)
+            return
+        # Clear spinner line, print text above, spinner redraws on next tick
+        self._write(f"\r\033[K  {text}", flush=True)
+
     def stop(self, final_message: str = None):
         self.running = False
         if self.thread:

run_agent.py

@@ -2362,6 +2362,24 @@ def run_conversation(
                 # Handle assistant response
                 if assistant_message.content and not self.quiet_mode:
                     print(f"{self.log_prefix}🤖 Assistant: {assistant_message.content[:100]}{'...' if len(assistant_message.content) > 100 else ''}")
+
+                # Notify progress callback of model's thinking (used by subagent
+                # delegation to relay the child's reasoning to the parent display).
+                # Guard: only fire for subagents (_delegate_depth >= 1) to avoid
+                # spamming gateway platforms with the main agent's every thought.
+                if (assistant_message.content and self.tool_progress_callback
+                        and getattr(self, '_delegate_depth', 0) > 0):
+                    _think_text = assistant_message.content.strip()
+                    # Strip reasoning XML tags that shouldn't leak to parent display
+                    _think_text = re.sub(
+                        r'</?(?:REASONING_SCRATCHPAD|think|reasoning)>', '', _think_text
+                    ).strip()
+                    first_line = _think_text.split('\n')[0][:80] if _think_text else ""
+                    if first_line:
+                        try:
+                            self.tool_progress_callback("_thinking", first_line)
+                        except Exception:
+                            pass
 
                 # Check for incomplete <REASONING_SCRATCHPAD> (opened but never closed)
                 # This means the model ran out of output tokens mid-reasoning — retry up to 2 times