Simple Linter AI Agent

A basic Node.js example demonstrating the "LLM + Loop + Tools" pattern for an AI agent. This agent uses an LLM (like GPT-4o Mini) and ESLint via a custom tool to automatically fix JavaScript linting errors in a target file.

Prerequisites

• Node.js (v18 or later recommended)
• npm (usually comes with Node.js)
• An OpenAI API Key

Setup

1. Clone the repository:

   git clone <your-repository-url>
   cd <repository-directory-name>

2. Install dependencies:

   npm install

3. Configure your API Key:
   • Create a file named .env in the root of the project.
   • Add your OpenAI API key to the .env file:

     OPENAI_API_KEY=sk-YourSecretOpenAiApiKeyGoesHere

Running the Agent

1. Prepare the target file: Ensure the file target_code.js exists in the root directory and contains JavaScript code with linting errors (according to the .eslintrc.json rules). You can modify this file or create your own.
2. Run the agent:

   node index.js

What it Does

• Reads the content of target_code.js.
• Calls the OpenAI API (configured model, e.g., gpt-4o-mini).
• Uses the runLinter tool, which executes ESLint to find errors.
• The LLM analyzes the errors and attempts to fix the code.
• It uses the runLinter tool again to verify the fixes.
• Prints the final corrected code to the console.
• Saves the corrected code to target_code_corrected.js.

Customization

• Target File: Change the TARGET_FILE_PATH variable in index.js to point to a different file.
• Linting Rules: Modify the .eslintrc.json file to adjust ESLint rules.
• LLM Model: Change the model value in src/config.js.
• System Prompt: Adjust the instructions in src/config.js (be careful, this affects agent behavior significantly).