fix: handle rate limiting with backoff, fix auth error detection

src/monarch_mcp_custom/auth.py

@@ -123,8 +123,15 @@ def retry_on_auth_error(max_retries: int = 1):
                         "401" in error_str
                         or "unauthorized" in error_str
                         or "authentication" in error_str
-                        or isinstance(e, (RequestFailedException, LoginFailedException))
+                        or isinstance(e, LoginFailedException)
                     )
+                    is_rate_limit = "429" in error_str or "rate limit" in error_str
+
+                    if is_rate_limit and attempt < max_retries:
+                        import asyncio
+                        logger.warning(f"Rate limited in {func.__name__}, waiting 15s before retry...")
+                        await asyncio.sleep(15)
+                        continue
 
                     if is_auth_error and attempt < max_retries:
                         logger.warning(