Customizing CAMEL-AI intelligences roles and tasks is mainly implemented through Python code as follows:
- Defining Intelligent Body Roles: Create an intelligent body using the ChatAgent class and set up system prompts
from camel.agents import ChatAgent
sys_msg = BaseMessage.make_assistant_message(role_name="Assistant", content="Your role description")
agent = ChatAgent(system_message=sys_msg, model=model) - task sth.: Trigger specific tasks with user messages
user_msg = "Your task instruction"
response = agent.step(user_msg) - Extended functionality: Toolset can be loaded for intelligences
from camel.toolkits import SearchToolkit
tools = [*SearchToolkit().get_tools()]
agent = ChatAgent(system_message=sys_msg, model=model, tools=tools)
The system supports a variety of predefined roles (e.g., researcher, analyst, educator, etc.) as well as fully customizable roles. By adjusting the content of system prompts, the behavioral patterns and areas of expertise of the intelligences can be finely controlled.
This answer comes from the articleCAMEL-AI: An Open Source Framework for Building Multi-Intelligent Collaborative SystemsThe
May not be reproduced without permission:AI productivity tools " How to customize the Intelligentsia roles and tasks in CAMEL-AI?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil