Skip to content

README.md

Latest commit

 

History

History
90 lines (62 loc) · 4.28 KB

README.md

File metadata and controls

90 lines (62 loc) · 4.28 KB

🤖AI Coding Agent:

An intelligent command-line utility for developers that enables seamless Python code creation, editing, execution, and file management using natural language, powered by the Gemini API's advanced function calling.

Python LLM Tools Framework Current Support

✨ Core Features & Technical Highlights

This agent uses a function-calling architecture to translate natural language commands into specific file system and execution operations, acting as a powerful developer utility.

Feature Description Technical Implementation
📝 Create File Generates a new Python script with code content specified via natural language. Utilizes the Gemini API for reliable code generation, ensuring proper syntax before writing to disk.
📖 Read File Displays the contents of an existing file back to the user for review. A dedicated system tool that reads the file and returns its content as observation.
✏️ Overwrite/Edit Modifies the content of an existing file based on instructions like "fix bug" or "add function." The agent sends the existing file content + the user's instruction to the LLM to output the complete, revised code.
▶️ Run File Executes the specified Python file safely within a controlled subprocess environment. Employs Python's built-in subprocess module (subprocess.run()) to capture and return the stdout (output) and stderr (errors).
⚡ Python-Only Focus The agent's prompt is heavily engineered to constrain its outputs strictly to valid Python 3.x logic, maximizing reliability. A strong system instruction is used to enforce the agent's persona, its available tools, and the single-language scope.

🛠️ Project Architecture

The application is structured into key components that manage the interaction loop between the user, the Gemini LLM, and the local file system tools.

File / Component Purpose Functionality
agent.py Core Execution Loop Initializes the Gemini model, manages the chat history, and orchestrates the tool-calling workflow.
tools.py Agent Tools Library Contains the actual Python functions (create_file, read_file, overwrite_file, run_file) that define the agent's capabilities.
requirements.txt Dependencies Lists all required Python libraries, primarily google-genai and related utilities.
.env Configuration Stores the required GEMINI_API_KEY for secure access to the LLM.

⚙️ Setup and Installation

Prerequisites

  1. Python 3.8+
  2. Gemini API Key: Obtain a key from Google AI Studio.
  3. Google-genai: To use function calling and Gemini API

Steps

  1. Clone the repository:

    git clone [YOUR-REPO-URL]
cd gemini-ai-coding-agent

  2. Create and activate a virtual environment:

    python -m venv venv
.\venv\Scripts\activate # Windows
# source venv/bin/activate # macOS/Linux

  3. Install dependencies:

    pip install -r requirements.txt

  4. Configure Environment Variables (.env file): Create a file named .env in the root directory and add your API key:

    GEMINI_API_KEY=your_gemini_api_key_here

🚀 How to Run the Application

The agent operates as a persistent, interactive chat session in your terminal, allowing for rapid iteration on code.

Start the Agent

Run this command in your terminal:

py agent.py

Credits and Thanks

Thanks to @freecodecamp and @boot.dev for providing the knowledge and guidance for this project