Goose: open source scalable programming intelligences that automate the full range of programming tasks

General Introduction

Goose is an open source AI agent tool developed by Block, Inc. designed to help developers automate everyday development tasks. Supporting a wide range of Large Language Models (LLMs) and interacting with users via the command line or desktop application interfaces, Goose performs the full range of tasks from code writing and editing to testing and deployment, dramatically improving development efficiency. The tool excels especially in task planning and execution, understanding complex development requirements and automating related tasks.

Extension tool: https://block.github.io/goose/v1/extensions/

Function List

• Code generation and editing: Automatically write or modify code based on user input.
• Task management and implementation: Create task schedules and automate them, including updating dependencies, running tests, and more.
• Integrated Development Environment Support: Integrate with development tools such as JetBrains IDE to provide a smoother development experience.
• Command line and desktop applications: Supports CLI and desktop user interfaces, adapting to different user habits.
• Extended Support: Integration with external services such as GitHub, Google Drive, etc. is supported through a plug-in system.
• file operation: Automatically handle file reads, writes, backups and version control.
• performance optimization: Automate code optimization and performance analysis.

Using Help

Installing Goose

Desktop application installation:

1.Download the installation package::

• Visit the official Goose GitHub page.
• Click on "Releases" to find the latest version for download.

2.Unzip the file::

• file the downloaded.zipThe file is extracted to a local folder.

3.Running the application::

• Double-click on the unzipped executable to start Goose.

CLI Installation:

1.Installation via curl::

curl -fsSL https://github.com/block/goose/releases/download/stable/download_cli.sh | sh

This command will automatically download and install Goose on your system.

Using Goose

Start Goose:

• CLI: Open a terminal and typegooseYou can access the interactive interface.
• desktop application: After launching the application, you will be presented with an easy-to-use interface.

Basic Operations:

• Configuring the LLM Provider::
  • The first time you use it, Goose will guide you through the process of selecting and configuring an LLM provider (e.g. OpenAI, Google Gemini, etc.), which requires you to enter the appropriate API key.
    • Run goose configure or configure via the settings menu of the desktop app.
• task order::
  • In the CLI, enter your development task or problem directly. For example

    G> create a new branch in this project and commit a file change
    

  • In the desktop application, type the task in the input box and press Enter Goose will start executing.
• Monitoring and interrupting tasks::
  • For long tasks, you can use CTRL+C or CTRL+D to interrupt or exit the current operation.

The function operates in detail:

• code editor: Goose can modify existing code to suit your needs. For example, you can ask Goose to migrate a function from Python 2 to Python 3.
  • importationG> Migrate this Python file to Python 3After that, Goose will browse the file and perform the necessary changes.
• Task automation: Suppose you need to update a dependency library and run a test:
  • importationG> Update the pytest library and run all tests. Goose looks for and updates the specified library and then executes the test commands.
• Extension and Integration::
  • To enable Goose integration with GitHub, after the configuration is complete, you can manage Pull Requests or Issue directly in Goose, for example

    G> View and merge PR #42
    

  • For Google Drive, you can set up Goose to handle documents

    G> Download the latest version of the document from Google Drive and update it locally
    

• version control: Goose handles Git operations automatically, but you can also explicitly instruct it to manage branches, commits, etc.

  G> create a new branch named feature/newLogin
  

Caveats:

• Ensure that your LLM provider account has sufficient API call credits to avoid interruptions during critical tasks.
• Update Goose regularly for the latest features and security patches.