Malcolm Bastien

Flappy Bird Development Workflow

Here’s a Mermaid.js flow diagram illustrating how task-master, Open Code (the environment I’m in), Zed Editor, Gemini (the underlying AI), and GitHub Pages collaborate in the development of the Flappy Bird clone, along with a summary explanation:

graph TD
    A[User] --> B{Zed Editor};
    B --> C;
    C --> D[Open Code / Agent Environment];
    A --> D;

    subgraph C[Project Files/Codebase]
        F{tasks.json & *.md files}
    end

    D -- "Executes commands via MCP" --> E[Task Master CLI];
    E -- "Manages tasks, plans" --> F;
    E -- "Leverages for AI capabilities" --> G(Gemini LLM);
    D -- "Leverages for AI capabilities" --> G;

    G -- "Provides task analysis, expansion, PRD parsing" --> E;
    F --> D;

    D -- "Reads/Edits code" --> C;
    C -- "Version Control" --> H[GitHub Repository];
    H -- "Hosts Game" --> I[GitHub Pages];

    subgraph Development Workflow
        A -- "Writes/Edits Code" --> C;
        A -- "Interacts with Agent" --> D;
        D -- "Automates Dev Tasks" --> E;
    end

    subgraph Deployment
        H -- "Publishes" --> I;
    end

Summary of the Development Workflow

This diagram illustrates the integrated development workflow for the Flappy Bird clone. At the center of the development is the User (A), who interacts with the Zed Editor (B) for direct code writing and editing, and with the Open Code / Agent Environment (D), which is where I, the AI assistant, operate. This environment is flexible and could be swapped out for other agent interfaces like Gemini CLI, Kilo Code, or Zed Editor’s built-in agent interface.

The Project Files/Codebase (C) is the central repository for all game code and assets. Changes from Zed Editor or automated edits from Open Code are reflected here.

Within the Open Code environment, the Task Master CLI (E) is a critical tool. This connection is enabled through MCP, allowing Open Code to execute task-master commands to manage the entire development process: parsing requirements from prd.txt, generating and expanding tasks, tracking their status in tasks.json and individual .md files (F), and analyzing complexity.

The intelligence behind both task-master’s advanced features (like task expansion and PRD parsing) and my own reasoning and code generation capabilities comes from the Gemini LLM (G). Gemini processes prompts and data to provide task-related insights and generate code or instructions.

As development progresses, the code in the Project Files/Codebase (C) is committed to a GitHub Repository (H). Finally, the completed game is published from the GitHub Repository to GitHub Pages (I), making the Flappy Bird clone accessible online. This setup creates a loop where requirements are broken down, tasks are managed, code is written and refined, and the project is deployed, all facilitated by AI-powered tools and platforms.