⚔️ AI vs AI Battle — Adjust the Attack AI and Defense AI IQ sliders to trigger a live cyber battle. A high Attack IQ unleashes advanced threats like zero-days and polymorphic malware; a low Defense IQ means the server gets breached. 📊 Anomaly Detection — The AI learns what normal traffic looks like, then flags deviations. Enable Attack Mode to mix in malicious requests, then use the Sensitivity slider to experience the precision-recall trade-off between catch rate and false alarms first-hand. 🎭 Deepfake Attack — A GAN-generated fake face goes up against an AI detector. Raise the Quality slider and pixel artifacts fade, making detection progressively harder. The GAN architecture that produces the deepfake is shown alongside. 🔓 LLM Jailbreak — Step through five real prompt injection techniques — direct request, roleplay framing, DAN persona override, language obfuscation, and authority framing. Lower the Guard Level and watch certain techniques punch through the safety layer. ☠️ Data Poisoning — Increase the poison rate in the training dataset and spam detection accuracy silently collapses. The model appears healthy but starts classifying spam as "Safe Email" — a textbook silent failure, visualised in real time. 🧠 Zero Trust AI — Select one of six users (normal employee, external IP, unusual login time, personal device, and more), press Analyze, and the AI evaluates five factors — IP reputation, access time, device trust, user role, and risk score — to grant or deny access on the spot.

A six-concept cybersecurity explorer — all concepts are visualised without any real attack code. 🔍 Reconnaissance — A rotating scan beam sweeps across the target network, probing for open ports (SSH, HTTP, MySQL, and more). Every discovered port is logged to the terminal in real time. 🔐 Brute Force — An automated tool fires thousands of password guesses per second. Lower the password strength slider and it cracks in moments; raise it and the time climbs dramatically. Real GPUs can attempt ~1 billion passwords per second. 🎣 Phishing — Renders a convincing fake inbox showing suspicious sender domains and urgency tactics, exactly as a victim would see them. Switch to the "Fake Website" tab to simulate a victim entering credentials into an attacker-controlled login page that looks identical to the real thing. 💉 SQL Injection — Toggle between "Normal Login" and "SQL Injection" to see how a single line — ' OR '1'='1 — bypasses the entire password check and exposes every user record in the database. The correct defense (Prepared Statements) is shown in code alongside. 🔒 Encryption — Alice encrypts a message with an AES key and sends it across the network. An attacker intercepts the packet mid-transit but sees only unreadable ciphertext. Click "SEND ENCRYPTED" to watch the full flow. Estimated crack times for 64, 128, and 256-bit keys are shown for comparison. 🛡️ Firewall & Defense — Live packets stream in from the internet and are either passed or blocked at the firewall. Add more rules with the slider and watch the block rate climb.