🛡️ BrainOps Guard

Making AI actually work inside companies by turning it into a controllable, safe, and context-aware execution engine.

---

⚠️ The Problem

AI is powerful, but inside companies, it often fails—not because the models are weak, but because the context is broken.

• Scattered Knowledge: Company info is trapped in docs, chats, and heads.
• Data Anxiety: Employees fear sharing sensitive data with LLMs.
• The Black Box: AI is generic and unreliable without specific company logic.

---

💡 The Solution

BrainOps Guard is an AI Extension Layer that sits between user input and any AI model (ChatGPT, Claude, Gemini, etc.). It acts as a governance wrapper that:

1. Context Builder: Converts messy documentation into structured behavioral blueprints.
2. Context Guard: Filters and redacts sensitive data (PII, tokens) before it hits the model.
3. Governance Layer: Enforces company-defined rules and rate limits in real-time.

---

🚀 Key Features

• 🧠 Intelligent Wrapper: Injects company-specific system prompts dynamically.
• 🛡️ PII/Sensitive Data Filter: Automatically detects and redacts sensitive information.
• 📎 Smart File Extraction: Upload policies (.txt, .pdf) and extract executable rules instantly.
• 🎛️ Developer Control Panel: Toggle "Safe Mode," adjust rule caps, and simulate prompt injection.
• 📊 Transparent Execution: See exactly how the AI decided (ALLOWED, BLOCKED, or SANITIZED).
• 🔌 Model Agnostic: Works with OpenAI, Groq, Ollama, or any AI provider.

---

🛠️ Tech Stack

• Frontend: React + Vite (Fast HMR, Modern build pipeline)
• Styling: Tailwind CSS v4 (Glassmorphic UI, Dark theme)
• Logic: Pure JavaScript (No heavy backend dependencies)
• Deployment: Ready for Vercel, Netlify, or as a self-hosted dashboard.

---

📥 Installation

Get BrainOps Guard running locally in seconds:

1. Clone the repository:

   git clone https://github.com/aarav214/BrainOps.git
   cd BrainOps/Extension

2. Install dependencies:

   npm install

3. Start the development server:

   npm run dev

4. Open in your browser: Navigate to http://localhost:5173

---

📖 How to Use

1. Build Context: Add rules manually in the "Configuration Panel" or upload a policy document.
2. Generate Skill: Click "Generate Skill" to compile your rules into a structured AI behavior blueprint.
3. Configure Governance: Toggle Safe Mode or Sensitive Data filters in the Control Panel.
4. Test Execution: Enter a user query in the "User Input Layer" and click Run with BrainOps.
5. Inject (Optional): Toggle "Inject into AI Tools" to see the final system prompt and copy it for use in external LLMs.

---

🏗️ Architecture

graph TD
    A[User Input] --> B[BrainOps Guard Wrapper]
    B --> C{Governance Check}
    C -->|Blocked| D[Decision: BLOCKED]
    C -->|Pass| E[Context Guard / Redaction]
    E --> F[System Prompt Injection]
    F --> G[AI Model Call]
    G --> H[Post-Process Result]
    H --> I[Decision: ALLOWED/SANITIZED]

Loading

---

🔮 Roadmap

• Real-time API Integrations: Connect directly to OpenAI/Groq APIs.
• Chrome Extension V3: Direct browser injection for ChatGPT/Claude.
• Persistence: Save rule-sets and history via LocalStorage/PostgreSQL.
• Advanced NLP: Vector-based rule matching for higher accuracy.

---

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

We don’t just give AI knowledge—we make it executable, controllable, and safe.