ScoutShield AI

ScoutShield AI is a defensive scam and phishing scanner that checks suspicious links, Discord/Telegram messages, giveaway text, login requests, and screenshot text before users click or share private information.

The app is now packaged as a complete deployable web product: a polished browser UI, a local explainable risk engine, safe training cases, and a Netlify serverless endpoint for live DNS, domain-age, and abuse-feed checks.

Features

• Link Checker: scores URLs with local phishing heuristics and optional live Netlify checks.
• Message Scanner: detects social-engineering patterns such as urgency, fake rewards, account verification, 2FA/code requests, wallet theft, and impersonation.
• Screenshot Scanner: previews uploaded screenshots and analyzes the visible text entered by the user.
• Explainable report card: returns a risk score, verdict, confidence level, key signals, category bars, and a recommended next action.
• Safe training lab: built-in non-malicious cases let you validate the scanner without using real harmful links.
• Privacy-first fallback: local scoring works even when serverless live checks are unavailable.

Project structure

.
├── index.html
├── styles.css
├── app.js
├── netlify.toml
├── package.json
├── netlify/functions/scan-url.js
├── TEST_CASES.md
├── DEPLOYMENT.md
└── SECURITY.md

Run locally

Open index.html directly in a modern browser for the local scanner.

For a local static server:

python3 -m http.server 8080

Then open:

http://localhost:8080

Local file/static-server mode uses browser-side scoring. Live DNS, RDAP, and abuse-feed checks run after deployment on Netlify.

Deploy on Netlify

1. Push this folder to a GitHub repository.
2. Create a new Netlify site from the repository.
3. Use the default configuration from netlify.toml:
   • publish directory: .
   • functions directory: netlify/functions
4. Deploy.
5. Test the Link Checker with a safe URL and confirm the report summary includes live-check details.

Technical approach

ScoutShield combines two layers:

1. Local explainable engine in app.js

   • suspicious TLD detection
   • URL shortener detection
   • brand lookalike patterns
   • reward bait and urgency detection
   • credential/2FA/seed phrase detection
   • impersonation and secrecy pressure detection
   • screenshot text scoring

2. Serverless live checks in netlify/functions/scan-url.js

   • HTTPS status
   • raw IP detection
   • DNS resolution through Google DNS-over-HTTPS
   • RDAP domain registration age lookup
   • URLHaus exact URL lookup

The scanner is intentionally transparent: every warning includes the signal that caused it and a student-friendly explanation.

Safety and ethics

ScoutShield AI is a defensive education and protection tool. It does not create phishing pages, steal credentials, bypass systems, or provide offensive instructions. Built-in training cases use safe example domains and non-malicious content.