🕵️‍♂️ Dark Pattern Detective

"The internet is full of traps. We built the detector."

Dark Pattern Detective is an AI-powered forensic tool designed to expose deceptive user interfaces. By leveraging Multi-Agent Systems (MAS), it combines computer vision and semantic text analysis to identify "Dark Patterns"—manipulative design choices intended to trick users into giving up privacy, money, or attention.

---

Demo :

🚩 The Problem

The modern web is plagued by Dark Patterns:

• Urgency Scarcity: "Only 1 room left!" (when there are 50).
• Confirmshaming: "No, I hate saving money."
• Visual Interference: Hiding the "Unsubscribe" button in grey text on a white background.

Traditional ad-blockers use rigid scripts. They fail when designers change a CSS class. We needed a solution that sees and thinks like a human.

🧠 Architecture & Agents

We utilize a Multi-Agent Architecture to deconstruct web pages.

graph LR
    A[User Input URL] --> B(🕵️ Navigator Agent)
    B --> C{Data Extraction}
    C -->|Screenshot| D[👁️ Vision Agent]
    C -->|DOM/Text| E[🧠 Analyst Agent]
    D --> F[⚖️ Risk Aggregator]
    E --> F
    F --> G[Final Safety Report]
    style A fill:#333,stroke:#fff,stroke-width:2px,color:#fff
    style G fill:#b92b27,stroke:#333,stroke-width:2px,color:#fff

Loading

The Agents

1. 🕵️ Navigator Agent: Uses Selenium/Playwright to browse the live site, handling popups and scrolling to capture the true user experience.
2. 👁️ Vision Agent: A Multi-modal LLM (e.g., Gemini Pro Vision / GPT-4V) that scans screenshots for "Visual Interference" and low-contrast traps.
3. 🧠 Analyst Agent: Parses the HTML text to detect manipulative wording, aggression, or guilt-tripping language.

---

🛠️ Tech Stack

• Core: Python 3.10+
• Orchestration: LangChain / CrewAI
• LLMs: Google Gemini Pro / OpenAI GPT-4
• Frontend: Streamlit / Gradio
• Computer Vision: PIL / OpenCV
• Browser Automation: Selenium / Playwright

---

🚀 Installation & Usage

Prerequisites: Ensure you have Python 3.10+ installed on your system.

1. Clone the Repository

git clone [https://github.com/B-Mohid/Dark_Pattern_Detective.git](https://github.com/B-Mohid/Dark_Pattern_Detective.git)
cd Dark_Pattern_Detective

2. Set up Virtual Environment

# For Linux/Mac (Fedora)
python3 -m venv venv
source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

4. Configure Environment Variables

Create a .env file in the root directory:

GOOGLE_API_KEY=your_gemini_key_here
# or
OPENAI_API_KEY=your_openai_key_here

5. Run the Detective

streamlit run app.py
# or if using Gradio
python app.py

---

🧪 Testing

To see the agent in full effect, test it against these educational "Honey Pot" sites:

• User In Yer Face (The ultimate stress test)
• Deceptive Design Hall of Shame

---

🗺️ Roadmap

• Basic Dark Pattern Detection (Text & Visual)
• Hugging Face Deployment
• Browser Extension: Real-time protection overlay.
• Auto-Fill Defense: Prevent agents from signing up for spam.
• Mobile App: Scan screenshots from your phone.

---

🤝 Contributing

Contributions are welcome! Please fork the repository and submit a pull request.

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

---

📄 License

Distributed under the MIT License. See LICENSE for more information.

---

Made with lots of 💖️ using Fedora 42 💫️ 🤖️