Hack_NCState_2026

Team RNW Hackers

Integrity OS - Layered Protection Extension

Integrity OS is a browser extension that assigns a real-time trust score to every webpage a user visits using a layered system. The extension doesn't just stop there, it extendes user defined protections: disabling checkout, blurring sensetive/hateful content, providing audio summaries of webpages, etc.

LAYER 0: USER PROFILE

One-time setup

User selects protection level:

• Minimal - Information only (for tech-savvy users). {Use gemini key to summarize findings}
• Scam Protection - Block suspicious payments
• Impulse Buy Protection - Cooling-off warnings
• Engagement Protection - Engagement bait warnings

---

LAYER 1: QUICK TRUST CHECK

< 1 second, runs on every page

Lightweight checks:

• ✓ Domain age lookup (WHOIS API)
• ✓ SSL certificate validation
• ✓ Known scam database check
• ✓ Blacklist comparison (import a database of known trusted websites like Amazon)
• ✓ Basic page structure analysis:
  • Excessive pop-ups?
  • Payment forms present?
  • Comment sections detected?
  • Suspicious keywords ("limited time", "act now")

OUTPUT: Trust Score (0-100) [Base trust score is null]

---

DECISION POINT

Score ≥ 60 (Trustworthy)     |     Score < 60 (Suspicious)
           ↓                 |                ↓
    ✓ PASS - Show badge      |       TRIGGER LAYER 2
                             |        Deep AI Scan

---

LAYER 2: DEEP AI SCAN

Only triggered if Score < 60

Expensive AI analysis:

🤖 GOOGLE GEMINI

• Extract all text from page
• Run AI generation detection
• Output: % AI-generated (0-100%)
• Flag suspicious patterns (urgency, scam language)

ELEVENLABS

• Allow the user to listen to an audio render of the summary that Gemini finds.

RESULT: Updated Trust Score incorporating AI findings

---

LAYER 3: PROTECTION RULES

Based on User Profile + Final Score

IF User has "Scam Protection" enabled:

• AND Final Score < 40
• AND Payment forms detected:
  • → HARD BLOCK payment submission
  • → Show detailed warning with all red flags
  • → [Notify trusted contact - out of scope]

IF User has "Impulse Buy Protection" enabled:

• AND Checkout flow detected:
  • → Show cooling-off prompt
  • → [24hr delay - mentioned but not built]

IF User has "Engagement Protection" enabled:

• AND Comment form on low-trust site:
  • → Show context warning before posting
  • → [Delay submission - mentioned but not built]

IF User has "Minimal" protection:

• → Just display trust score badge
• → Show AI detection results if available
• → No blocking, full user autonomy

---

LAYER 4: USER INTERFACE

Visual feedback based on final score:

Score Range	Badge	Status
80-100	🟢	Verified Safe
60-79	🟡	Proceed with Caution
0-59	🔴	Threat Detected

Click badge to see full report:

• Trust score breakdown
• AI detection results
• Specific red flags
• Recommended actions