Hugging Face releases a barebones library for agents (huggingface.co)

The multi-step agent API they've produced feels so much like reinforcement learning. Neat!

    memory = [user_defined_task]
    while llm_should_continue(memory): # this loop is the multi-step part
        action = llm_get_next_action(memory) # this is the tool-calling part
        observations = execute_action(action)
        memory += [action, observations]