QR codes bypassed every URL safety habit we spent a decade…

QR codes bypassed every URL safety habit we spent a decade building

Somewhere around 2010, internet security education settled on a reasonable heuristic: before clicking a link, check where it actually goes. Hover over it, read the URL, make sure it's not typosquatting the real domain. That instinct, trained over years, is completely useless against a QR code.

A QR code gives you nothing to inspect. You point your phone at it, and the camera's URL preview is a single line that disappears in under two seconds — if you see it at all. Most people scan and open without pausing to read. The gap between "that habit worked" and "this attack surface didn't exist yet" is where quishing — QR phishing — lives.

The attack is cheap and physical: print a sticker with your malicious QR code and place it over the legitimate one. Parking meters, restaurant menus, package tracking texts, conference room equipment instructions — all use QR codes now, all became valid targets after 2020. The city of Austin had to issue a public warning after fake payment stickers appeared on parking meters. The FBI issued an advisory in 2022. The volumes since then have only gone up.

The email vector is different but the same gap applies. Email filters trained for years to catch suspicious URLs and domains. Embedding that URL as a QR code inside a PDF attachment routes around every link scanner. The attachment isn't a link. The QR code is an image. The scanner's job ends before the attack begins.

There's an architectural mismatch here that nobody has cleanly solved. Browsers will warn you about known phishing domains; they can't warn you about a QR code you scanned before the browser loaded. Antivirus doesn't scan physical stickers. Certificate transparency logs the domains but not whether a QR code encoded one of them.

Does anyone actually read the URL after scanning a QR code? Or is "scan and trust" just where we landed?

0 replies