ShieldCore WAF is an AI-powered Web Application Firewall dashboard built to detect, analyze, and neutralize web-layer threats in real time. Trained on the CSIC 2010 dataset (61,000 samples), ShieldCore combines a 5-stage neural inference pipeline with live global threat mapping, attacker behavioral profiling, and a one-of-a-kind Attacker vs Defender Mirror — giving security teams true visibility into how attacks are constructed and crushed.
Built in 24 hours at a hackathon. Zero compromises on depth or design.
| Metric | Value |
|---|---|
| 📦 Dataset Size | 61,000 samples (CSIC 2010) |
| 🚫 Total Blocks Executed | 125,026+ |
| 📡 Packets Inspected | 2,846,430+ |
| 🌍 Global Attacks Tracked | 25,124 across 42 regions |
| ⚡ Avg Inference Time | 5.2ms |
| 🎯 Model Accuracy | 94.3% |
| 📶 Ingress Velocity | 482.4k packets/min |
The core intelligence engine of ShieldCore. Submits raw HTTP payloads through a 5-stage neural pipeline:
INGRESS RECEIVED → URL DECODE → BASE64 DECODE → TOKENIZATION → NEURAL CLASSIFY
- Detects SQL Injection, XSS, Path Traversal, and anomalous patterns
- Outputs Attack Class, Severity, Confidence Score, and Inference Time
- Displays an AI Forensic Verdict with plain-language explanation
- Supports RAW / DECODED payload view toggle
- OWASP classification + Export Intelligence as JSON or PDF
- Threat library with 61k corpus (59% valid / 41% anomalous)
Edge-wide telemetry across global ingress nodes, streamed in real time.
- Tracks Total Ingress, Neutralized Threats, and Avg Threat Index live
- Color-coded entries:
BLOCKED/SAFEwith intelligence match scores - Search by IP address or Attack Class
- Suspend/resume feed on demand
- Each entry shows IP, endpoint, HTTP method, latency, and threat verdict
ShieldCore's most powerful and unique module — a split-screen battle view between attacker and defender.
Attacker Side — Intercepted Hacker Workstation:
- Shows the full Payload Construction Stack step by step
- Start with Malicious Intent → Add Encoding → Obfuscate → Launch
Defender Side — ShieldCore Neural Ingress Core:
- Edge Node Intercept → De-obfuscation → Neural Transformer Activation → Final Decision
- Surfaces Forensic Evidence: decoded SQL injection patterns, boolean logic detection, OWASP A03:2021 violations
- Provides Defender Recommendations with one-click IP range blocking
- Scoreboard: Attacker vs Defender (0 vs 9 — defender wins 🛡️)
- Modes: Explain Mode, Live Intercept, Suspicious Anomaly
Strategic command center for long-term threat intelligence.
- Attack Vector Distribution — concentric ring chart of 25k anomalous samples
- Threat Intensity by Country — ranked bar chart (Russia, China, Iran, North Korea, India...)
- Attacker Sophistication Profile — rainbow arc chart profiling threat actor skill levels
- Real-Time Ingress Velocity — live line chart at 482.4 kbps
- Global Neural Ingress Map — world map with animated attack origin arcs
- Filter by: SQLi, XSS, DDoS, Malware, Phishing, APT, Botnet
- Search by Host Origin | Filter by Skill Level
The Global Neural Ingress map visualizes active attack trajectories across the world in real time using the CSIC 2010 simulation dataset.
- 25,124 total attacks across 42 affected regions
- Most active threat: Moscow / SQL Injection
- Threat Level: CRITICAL
- Animated dashed arc lines per attack vector with cluster counts per region
| Layer | Technology |
|---|---|
| Frontend | React / Next.js |
| Styling | Tailwind CSS |
| AI/ML | Neural Classifier (CSIC 2010) |
| Dataset | CSIC 2010 HTTP Dataset |
| Deployment | Vercel |
| Charts | Recharts / D3.js |
| Maps | Leaflet / MapBox |
shieldcore-waf/
├── app/
│ ├── page.tsx # Landing page
│ ├── analyzer/ # Payload Analysis Engine
│ ├── live-feed/ # Traffic Stream
│ ├── attacker-mirror/ # Attacker vs Defender
│ └── analytics/ # Global Analytics Dashboard
├── components/
│ ├── ThreatMap.tsx
│ ├── PayloadPipeline.tsx
│ ├── LiveFeedStream.tsx
│ └── AttackerMirror.tsx
├── lib/
│ └── neural-classifier.ts
└── public/
| Attack Type | Detection |
|---|---|
| SQL Injection | ✅ |
| Cross-Site Scripting (XSS) | ✅ |
| Path Traversal | ✅ |
| DDoS Patterns | ✅ |
| Malware Signatures | ✅ |
| Phishing Payloads | ✅ |
| APT Indicators | ✅ |
| Botnet Traffic | ✅ |
| Base64 Obfuscation | ✅ |
| OWASP Top 10 Coverage | ✅ |
🎯 24-Hour Hackathon Project Built under pressure, shipped with pride. This project represents what's possible when a focused team combines AI, security engineering, and design in a single sprint.
MIT License © 2025 ShieldCore Team
ShieldCore WAF v1.0.0 | CSIC 2010 | Strategic Defense Node Active 🛡️
"The best firewall doesn't just block threats — it understands them."