Prompt Injection Detector

Overview

This project is a web-based tool designed to detect and block possible prompt injection attacks targeting Large Language Models (LLMs). It utilizes HTML, Tailwind CSS, and JavaScript to provide a clean and interactive interface. The system scans user inputs for potentially malicious instructions before forwarding them to the LLM API.

Features

• Prompt Injection Detection: Uses the PromptFoo API to analyze user input and identify potential prompt injection attacks.
• Safe Input Handling: If an injection is detected, the request is blocked, and an alert is shown.
• Interactive UI: Built with Tailwind CSS for a responsive and modern design.
• LLM Querying: If the input is safe, the request is sent to Google's Gemini LLM API for processing.

Tech Stack

• Frontend: HTML, Tailwind CSS
• Backend: JavaScript (Client-side)
• APIs Used:
  • PromptFoo (for prompt injection detection)
  • Google Gemini LLM API (for generating responses)

Installation and Setup

1. Clone the repository:

   git clone https://github.com/your-repo/prompt-injection-detector.git
   cd prompt-injection-detector

2. Open index.html in a browser.
3. Ensure an internet connection to access the external APIs.

How It Works

1. User Input: The user types a prompt into the input box.
2. Injection Check: The input is sent to the PromptFoo API for analysis.
3. Result Handling:
   • If an injection is detected, the request is blocked, and an alert is shown.
   • If safe, the prompt is forwarded to the Google Gemini API.
4. Response Display: The AI-generated response is displayed on the webpage.

File Structure

/
│── index.html      # Main HTML structure
│── script.js       # JavaScript logic for API calls and validation
│── Assets/         # Directory for images and other assets

Usage

• Open the webpage.
• Enter a query in the text box.
• Click the send button.
• If the input is safe, you’ll receive an AI-generated response; otherwise, a warning will be displayed.

Future Improvements

• Add backend support for improved security.
• Enhance detection methods using custom AI models.
• Improve UI/UX with better feedback mechanisms.

Contributors

• Anish Raj Pandey
• Oasis Pandey
• Niraj Bhattarai

License

This project is open-source and available under the MIT License.