sentinel-scam-honeypo / docs /DEMO_SCRIPT.md
avinash-rai's picture
feat: GUVI final submission pack (docs, dashboard, telemetry)
7b1aabb
|
raw
history blame
2.35 kB
# πŸŽ₯ GUVI Demo Video Script (1 Minute)
**Goal:** Convince judges that this is a working, improved, and enterprise-ready honeypot system.
---
## 🎬 Scene 1: Introduction (0:00 - 0:10)
* **Visual:** Show the `README.md` with the "Sentinel Scam Honeypot" banner and badges.
* **Voiceover:** "Hi, this is Team [Team Name]. We've built Sentinel, an autonomous AI honeypot that traps scammers and extracts critical intelligence for Indian law enforcement."
## 🎬 Scene 2: The Attack (0:10 - 0:25)
* **Visual:** Split screen. Left: Your API/Dashboard. Right: Swagger UI (`/docs`).
* **Action:** Send a simulated scam message via Swagger UI (`/api/v1/analyze`).
* **Input:** `"Congratulations! You won 1 crore lottery! Send tax amount to verify at winner@paytm or call +91-9876543210"`
* **Voiceover:** "Here, we simulate a typical lottery scam message attacking our citizen hotline."
## 🎬 Scene 3: The AI Response (0:25 - 0:40)
* **Visual:** Show the JSON response in Swagger. Highlight:
* `scam_detected: true`
* `honeypot_response`: (e.g., "Wah! Sach mein?")
* `extracted_intelligence`: Phone number and UPI ID.
* **Voiceover:** "Our hybrid AI instantly detects the scam, identifying it as a 'Lottery Fraud'. It generates a believable response using our 'Sharma Uncle' persona to hook the scammer, while automatically extracting their UPI ID and phone number."
## 🎬 Scene 4: Enterprise Intelligence (0:40 - 0:50)
* **Visual:** Switch to the **Streamlit Dashboard** (`dashboard.py`). Show the graphs.
* **Action:** Point to "Top Scams" or "Live Threat Feed".
* **Voiceover:** "All data is aggregated in real-time. We cluster campaigns and generate threat intelligence reports compatible with CERT-In and NPCI formats."
## 🎬 Scene 5: Conclusion (0:50 - 1:00)
* **Visual:** Show the `ARCHITECTURE.md` diagram or a slide with "1.4 Billion Scale".
* **Voiceover:** "Sentinel is built to scale for 1.4 billion citizens using an asynchronous agentic architecture. We don't just block calls; we waste scammer time and stop the money flow."
---
## πŸ› οΈ Validation Checklist for Video
1. [ ] Are badges visible? (Shows polish)
2. [ ] Is the JSON output readable? (Proofs it works)
3. [ ] Is the Dashboard active? (Visual appeal)
4. [ ] Did you mention "Citizen Safety"? (Emotional hook)