CVE-2026-33306: bcrypt on JRuby is broken at cost=31
I found my first CVE. A signed integer overflow in bcrypt-ruby's Java backend causes cost=31 to skip all 2 billion key-strengthening rounds. The hash looks valid but protects nothing.
Howya.
I’m Kev. A software + DevOps engineer, traveler, Bitcoiner, and builder of fast, secure, low-maintenance things.
This site is an unapologetic home for my passions: projects, writing, adventures and whatever else fits.
I found my first CVE. A signed integer overflow in bcrypt-ruby's Java backend causes cost=31 to skip all 2 billion key-strengthening rounds. The hash looks valid but protects nothing.