The problem

Phishing is the payload that turns one distracted click into an incident. Most people forward a suspicious email to IT and wait — the goal here was to put a fast, private first-pass triage directly in the browser.

The approach

1. 01Analyzes across four areas: sender authentication (spoofing, domain reputation), content (urgency and social-engineering language), technical indicators (disguised URLs, domain age, SSL), and structural patterns (header anomalies, attachment risk).
2. 02Google Gemini 2.0 performs the analysis behind a clean threat score (0–100) with a named risk level: Critical, High, Medium, Low, or Safe.
3. 03Every result ships with the red flags it found and an immediate action recommendation, so the output teaches while it triages.
4. 04Deployed serverless on Google Cloud Run behind a GitHub Pages front end, returning assessments in under five seconds.
5. 05Zero-retention architecture: no storage, no logging, no cookies — the email text exists only for the duration of the request.

Demonstrates

• Phishing analysis
• Applied AI in a security workflow
• Cloud deployment (Cloud Run, GCP)
• Privacy-first system design