01 — Headers
Authentication Check
We parse the Authentication-Results header to read SPF (sending server authorization), DKIM (message integrity), and DMARC (domain policy alignment). Failures on all three are the single strongest signal that a message is spoofed.
02 — URLs
Link Inspection
Every URL is extracted and checked for four things: mismatched display text, URL shorteners hiding the destination, punycode/IDN lookalike domains, and suspicious paths (login, verify) on domains unrelated to the sender.
03 — Content
Red Flag Detection
Urgency language, credential-prompt phrases, attachment references, and generic greetings are all pattern-matched in the body. These are the psychological pressure points attackers exploit — we surface them plainly.