refactor(commands): 修改未知命令处理逻辑为返回None

移除orchestrator/agent.py中不再需要的命令检测逻辑，因命令处理已提前在bot/commands.py中完成

bot/commands.py

@@ -56,7 +56,7 @@ async def handle_command(user_id: str, text: str) -> Optional[str]:
     elif cmd in ("/help", "/h", "/?"):
         return _cmd_help()
     else:
-        return f"Unknown command: {cmd}\n\n{_cmd_help()}"
+        return None
 
 
 async def _cmd_new(user_id: str, args: str) -> str:

orchestrator/agent.py

@@ -55,12 +55,6 @@ Guidelines:
 MAX_ITERATIONS = 10
 _TOOL_MAP = {t.name: t for t in TOOLS}
 
-COMMAND_PATTERN = re.compile(r"^/(new|list|close|switch|retry|help)", re.IGNORECASE)
-
-
-def _looks_like_command(text: str) -> bool:
-    return bool(COMMAND_PATTERN.match(text.strip()))
-
 
 class OrchestrationAgent:
     """Per-user agent with conversation history and active session tracking."""
@@ -108,8 +102,8 @@ class OrchestrationAgent:
         logger.info(">>> user=...%s  conv=%s  msg=%r", short_uid, active_conv, text[:80])
         logger.debug("    history_len=%d", len(self._history[user_id]))
 
-        # Passthrough mode: if active session and not a command, bypass LLM
-        if active_conv and not _looks_like_command(text):
+        # Passthrough mode: if active session, bypass LLM (bot commands handled earlier)
+        if active_conv:
             try:
                 reply = await manager.send(active_conv, text, user_id=user_id)
                 logger.info("<<< [passthrough] reply: %r", reply[:120])