🛡️ TrustShield Sentinel

Privacy-first Insider Threat Detection for Wema Bank Hackathon (Challenge Five)

---

🌍 Overview

Insider threats are one of the toughest challenges in banking.

• Some employees may accidentally expose data.
• Others may intentionally misuse their access.
• Traditional monitoring tools often feel like “Big Brother”, creating mistrust and harming workplace culture.

TrustShield Sentinel solves this by detecting unusual staff activity with simple behavioral signals (not surveillance) and engaging employees with friendly nudges instead of accusations.

Employees become partners in security, not suspects.

---

🎯 Problem Statement

How might we detect and contain insider threats in Wema Bank while preserving employee trust and respecting privacy?

---

💡 Our Solution

TrustShield Sentinel monitors patterns, not people.

• ✅ Detects unusual logins, device use, or file activity.
• ✅ Sends supportive prompts to staff:

  “We noticed some unusual activity. Was this expected?”

• ✅ Preserves privacy by anonymizing data until a credible risk emerges.
• ✅ Provides a dashboard for security teams with risk signals (low/med/high) without naming employees unless escalation is needed.

---

🔎 How It Works

1. Detection (Backend)

   • Define role baselines (e.g., teller vs. HR).
   • Flag anomalies like:
     • Login after 10 pm.
     • Bulk file downloads.
     • New device or location.

2. Engagement (Frontend)

   • Employees receive nudges through modal popups or chatbot prompts.
   • Options:
     • ✅ Yes, this was me
     • 🚩 No, escalate
     • 💬 Need help

3. Privacy Safeguards

   • Only metadata (time, count, location) — no content.
   • Employee IDs are hashed/anonymized in dashboard.
   • Alerts auto-expire to prevent over-monitoring.

4. Security Dashboard

   • Shows anonymized risk alerts.
   • Risk heatmap or cards by level.
   • Drill-down only after multiple anomalies.

---

🏗️ Tech Stack

• Frontend: React (dashboard + employee modals)
• Backend: Node.js (Express) with mock MongoDB/JSON data
• Prototype Data: Sample employee logs with login times, devices, actions

---

🚀 Hackathon MVP Scope

• Simple anomaly rules engine (3–4 rules).
• Mock employee activity logs (10–15 users).
• Frontend dashboard with anonymized alerts.
• Modal/chatbot flow for employee confirmation.
• Pitch deck with story, benefits, and future potential.

---

---

🔮 Future Potential

• Expand from simple rules → AI anomaly detection.
• Integrate with IAM tools (Okta, Azure AD).
• Add gamified trust scores to build a security culture.
• Deploy as SaaS for banks & enterprises.

---

📜 License

MIT License (for hackathon demo purposes).