diff --git a/.gitignore b/.gitignore
index c6a733e..b286f93 100644
--- a/.gitignore
+++ b/.gitignore
@@ -96,3 +96,4 @@ tests/test_*.py
 loot/token_usage.json
 mcp_examples/kali/mcp_servers.json
 pentestagent/mcp/mcp_servers.json
+loot/notes.json
diff --git a/pentestagent/interface/tui.py b/pentestagent/interface/tui.py
index f79d9f3..caed3bc 100644
--- a/pentestagent/interface/tui.py
+++ b/pentestagent/interface/tui.py
@@ -648,34 +648,69 @@ class ThinkingMessage(Static):
 class ToolMessage(Static):
     """Tool execution message"""
 
-    # Standard tool icon and color (pa theme)
-    TOOL_ICON = "$"
-    TOOL_COLOR = "#9a9a9a"  # spirit gray
+    TOOL_COLOR = "#9a9a9a"
+    ARG_COLOR = "#6b6b6b"
+    HINT_COLOR = "#6b6b6b"
+
+    CHEVRON_COLLAPSED = "▶"
+    CHEVRON_EXPANDED = "▼"
+    HINT_TEXT = " (click to see result)"
+
+    expanded: bool = reactive(False, layout=True)
 
     def __init__(self, tool_name: str, args: str = "", **kwargs):
         super().__init__(**kwargs)
         self.tool_name = tool_name
         self.tool_args = args
+        self._result_widget: ToolResultMessage | None = None
 
     def render(self) -> Text:
         text = Text()
-        text.append(f"{self.TOOL_ICON} ", style=self.TOOL_COLOR)
-        text.append(f"{self.tool_name}", style=self.TOOL_COLOR)
-        text.append("\n", style="")
 
-        # Wrap args
+        chevron = (
+            self.CHEVRON_EXPANDED if self.expanded else self.CHEVRON_COLLAPSED
+        )
+
+        # Header line
+        text.append(f"{chevron} ", style=self.TOOL_COLOR)
+        text.append(self.tool_name, style=self.TOOL_COLOR)
+
+        # Hint text (only when result exists and is collapsed)
+        if self._result_widget and not self.expanded:
+            text.append(self.HINT_TEXT, style=self.HINT_COLOR)
+
+        text.append("\n")
+
+        # Tool arguments
         if self.tool_args:
             for line in wrap_text_lines(self.tool_args, width=110):
-                text.append(f"  {line}\n", style="#6b6b6b")
+                text.append(f"  {line}\n", style=self.ARG_COLOR)
 
         return text
 
+    def attach_result(self, result_widget: "ToolResultMessage") -> None:
+        """Attach a ToolResultMessage widget below this message."""
+        if self._result_widget is not None:
+            return
+
+        self._result_widget = result_widget
+        self._result_widget.display = self.expanded
+
+        # Mount directly after this widget
+        self.mount(self._result_widget, after=self)
+
+    def on_click(self) -> None:
+        self.expanded = not self.expanded
+        if self._result_widget:
+            self._result_widget.display = self.expanded
+
 
 class ToolResultMessage(Static):
     """Tool result/output message"""
 
     RESULT_ICON = "#"
-    RESULT_COLOR = "#7a7a7a"
+    RESULT_COLOR = "#124670"
+    OUTPUT_COLOR = "#17606d"
 
     def __init__(self, tool_name: str, result: str = "", **kwargs):
         super().__init__(**kwargs)
@@ -684,13 +719,14 @@ class ToolResultMessage(Static):
 
     def render(self) -> Text:
         text = Text()
+
         text.append(f"{self.RESULT_ICON} ", style=self.RESULT_COLOR)
         text.append(f"{self.tool_name} output", style=self.RESULT_COLOR)
-        text.append("\n", style="")
+        text.append("\n")
 
         if self.result:
             for line in wrap_text_lines(self.result, width=110):
-                text.append(f"  {line}\n", style="#5a5a5a")
+                text.append(f"  {line}\n", style=self.OUTPUT_COLOR)
 
         return text
 
@@ -1685,14 +1721,14 @@ class PentestAgentTUI(App):
     def _add_thinking(self, content: str) -> None:
         self._add_message(ThinkingMessage(content))
 
-    def _add_tool(self, name: str, action: str = "") -> None:
-        self._add_message(ToolMessage(name, action))
+    def _add_tool(self, name: str, action: str = "") -> ToolMessage:
+        tool_message = ToolMessage(name, action)
+        self._add_message(tool_message)
+        return tool_message
 
-    def _add_tool_result(self, name: str, result: str) -> None:
+    def _add_tool_result(self, tool_message: ToolMessage, name: str, result: str) -> None:
         """Display tool execution result"""
-        # Hide tool output - LLM will synthesize it in its response
-        # This prevents duplication and keeps the chat clean
-        pass
+        tool_message.attach_result(ToolResultMessage(name, result))
 
     def _show_system_prompt(self) -> None:
         """Display the current system prompt"""
@@ -3215,6 +3251,8 @@ Be concise. Use the actual data from notes."""
 
             from ..agents.base_agent import AgentState
 
+            last_tool_message: ToolMessage
+
             async for response in self.agent.agent_loop(task):
                 if self._should_stop:
                     self._add_system("[!] Stopped by user")
@@ -3241,7 +3279,7 @@ Be concise. Use the actual data from notes."""
                     for call in response.tool_calls:
                         # Show all tools including finish
                         args_str = str(call.arguments)
-                        self._add_tool(call.name, args_str)
+                        last_tool_message = self._add_tool(call.name, args_str)
 
                 # Show tool results
                 if response.tool_results:
@@ -3251,11 +3289,11 @@ Be concise. Use the actual data from notes."""
                             continue
 
                         if result.success:
-                            self._add_tool_result(
+                            self._add_tool_result(last_tool_message, 
                                 result.tool_name, result.result or "Done"
                             )
                         else:
-                            self._add_tool_result(
+                            self._add_tool_result(last_tool_message, 
                                 result.tool_name, f"Error: {result.error}"
                             )