perf(cli): defer AIAgent import to cut cold-start latency

cli.py

@@ -13,6 +13,8 @@ Usage:
     python cli.py --list-tools             # List available tools and exit
 """
 
+from __future__ import annotations
+
 import logging
 import os
 import shutil
@@ -477,7 +479,6 @@ from rich.text import Text as _RichText
 import fire
 
 # Import the agent and tool systems
-from run_agent import AIAgent
 from model_tools import get_tool_definitions, get_toolset_for_tool
 
 # Extracted CLI modules (Phase 3)
@@ -2029,6 +2030,8 @@ class HermesCLI:
         Returns:
             bool: True if successful, False otherwise
         """
+        from run_agent import AIAgent
+
         if self.agent is not None:
             return True
 
@@ -4056,6 +4059,8 @@ class HermesCLI:
         turn_route = self._resolve_turn_agent_config(prompt)
 
         def run_background():
+            from run_agent import AIAgent
+
             try:
                 bg_agent = AIAgent(
                     model=turn_route["model"],