This project features a Coder Agent designed to enhance developer productivity by automating code writing, an iterative code review feedback loop, and terminal command execution. The agent leverages tools to streamline workflows, ensuring accuracy and efficiency.
To get started, clone the repository and install the required dependencies:
git clone <repository-url>
cd <repository-directory>
pip install -r requirements.txtRun the main script to start interacting with the Coder Agent:
python agent.pyThe codebase includes the following agents:
- Coder Agent: Uses tools like
ShellTool,create_directory,find_file,create_file,update_file, andtavily_web_search. - Reviewer Agent: Uses tools like
get_files_in_directory,find_file,read_file, andtavily_web_search.
The agent uses a variety of tools to perform its tasks:
- ShellTool: Executes shell commands.
- create_directory: Creates a new directory.
- find_file: Searches for a file.
- create_file: Creates a new file.
- update_file: Updates an existing file.
- tavily_web_search: Performs web searches using the Tavily API.
- get_files_in_directory: Retrieves a list of files in a directory.
- read_file: Reads the contents of a file.
You can add new tools to match your specific workflow needs. For example, to add a tool for creating a React app with Vite, you can define it in tools.py and include it in the coder_tools list.
- Iterative Code Review Feedback Loop: Identify potential bugs and suggest improvements, then send it back to the code for implementation.
- Context-Aware Assistance: Offer suggestions based on project structure.
- Collaboration: Facilitate team collaboration with automated code reviews.
- CI/CD Integration: Automate testing and deployment processes.
- AI-Driven Learning: Personalize assistance based on coding patterns.
Contributions are welcome! Please submit a pull request or open an issue to discuss your ideas.
This project is licensed under the MIT License. See the LICENSE file for details.