This implementation adds a new get_available_models() method to the DebugClient class, which retrieves the list of available model configurations from the API and returns a list of their names. The send_message() method then calls this method and checks if the provided model_config_name is in the list of available models. If it's not, a ValueError is raised with an appropriate error message.

we can modify the `__init__` method to include the line `self.available_models = self.get_available_models()` so that the list of available models is retrieved only once during the instantiation of the DebugClient object: