autogpt AGENTS.md

1. 🎩 ManagerGPT

ManagerGPT serves as the orchestrator of your project, directing the other agents to execute tasks based on your input. When you provide a project prompt, ManagerGPT divides it into tasks for BackendGPT, FrontendGPT, DesignerGPT, and ArchitectGPT.

How ManagerGPT Works?

Let's say you want to develop a full-stack app that fetches today's weather in Python using FastAPI. ManagerGPT simplifies this process by breaking it down into specific tasks for each specialized agent:

• ArchitectGPT: ManagerGPT instructs ArchitectGPT to design the application's structure, encompassing both backend and frontend components, utilizing Python and FastAPI:

[*] "ManagerGPT": Executing task: "Develop a full stack app that fetches today's weather in python using FastAPI."

ManagerGPT articulates the project goal and communicates it to the ArchitectGPT through Gemini API, which then generates specific steps for architecting the application:

[*] "ArchitectGPT": Executing tasks: Task { description: "- Design the user interface for the weather app, including input fields for location and a display area for weather details.\n- Implement a function to fetch current weather data from a weather API in Python.\n- Create a FastAPI endpoint that calls the weather fetching function and returns the weather data in JSON format.\n- Integrate the FastAPI endpoint with the front end to display the fetched weather data on the user interface.\n- Handle error scenarios and provide appropriate user feedback.", scope: None, urls: None, frontend_code: None, backend_code: None, api_schema: None }

• DesignerGPT: ManagerGPT guides DesignerGPT in crafting a user-friendly interface tailored for presenting the weather forecast.

[*] "DesignerGPT": Executing tasks: Task { description: "- Utilize FastAPI in Python to develop a user interface for the weather app, integrating a location input field and display section for weather data.\n- Step 1: Define the layout and structure of the user interface, ensuring it's user-friendly and visually appealing.\n- Step 2: Implement the location search functionality, enabling users to input their desired location and retrieve relevant weather information.", scope: None, urls: None, frontend_code: None, backend_code: None, api_schema: None }

• BackendGPT: ManagerGPT assigns BackendGPT to implement the backend logic using FastAPI, responsible for fetching weather data from external sources.

[*] "BackendGPT": Executing tasks: Task { description: "- Using FastAPI in Python, create a backend for a weather application featuring a user interface with a location input field and weather data display section.\n- Step 1: Design and develop the user interface, ensuring it's user-friendly and visually appealing.\n- Step 2: Implement the location search functionality, enabling users to input their desired location and retrieve relevant weather information.", scope: None, urls: None, frontend_code: None, backend_code: None, api_schema: None }

• FrontendGPT: ManagerGPT directs FrontendGPT to develop the frontend interface, enabling users to interact with and visualize the weather data.

[*] "FrontendGPT": Executing tasks: Task { description: "- Using FastAPI in Python, create a user interface for a weather application.\n- Step 1: Create a user-friendly layout for displaying weather information.\n- Step 2: Implement FastAPI endpoints to fetch and display weather data.", scope: None, urls: None, frontend_code: None, backend_code: None, api_schema: None }

Finally, ManagerGPT aggregates the outcomes from all GPTs, consolidating the generated tasks into a coherent project plan. Here's the summary of the completed tasks: