
PhishShield AI is a real-time phishing intelligence tool that goes beyond "safe or dangerous." It tells you why something is dangerous, who is attacking you, what they wanted, and what you should do right now. Paste any suspicious email or URL and get back: A Threat Score (0–100) with visual gauge A Verdict (Safe / Suspicious / Dangerous) Specific Red Flag Indicators — not vague warnings, but precise findings like "sender domain paypa1.com mimics paypal.com using character substitution" The Attack Technique — credential harvesting, spear phishing, CEO fraud, etc. The Attacker's Goal — what data, money, or access they were trying to steal User Action — concrete steps you should take right now One thing worth flagging: this description says Llama 3.1 via OpenRouter, but the slide deck I built lists Gemini 2.5 Flash (from the original brief). You'll want to pick one and make them consistent — which AI model are you actually using? I can update the deck's tech stack slide to match.
31 May 2026