Dear mailop list, I am maintaining an open source tool called badkeys that identifies cryptographic public keys affected by known vulnerabilities.
I have used that to scan various collections of public keys, and I thought it'd be interesting for this community to share highlights from my DKIM scans. I already did some preliminary scans in 2024 [2] and a much more extensive scan lately. Stats here: https://monitor.badkeys.info/dkim/2026-08-11-dkim-badkeys.html Notable: * To this day, a relatively large number of DKIM keys are affected by the 2008(!) Debian OpenSSL bug. This likely stems from the fact that DKIM was created in 2007, within the timeframe when this bug was affecting Debian and was not patched yet, and people tend to leave DNS records alone once they created them. * A non-trivial number of DKIM records contain a private instead of a public key. (Likely, any signature generated for such a selector will not be validated correctly.) * Plenty of DKIM keys contain too short RSA keys (RSA 512 and smaller is breakable with open source tools on a modern-day laptop within days or even hours, 768 bit is breakable if you have a large cluster, but it's a challenge, 1024 bit is believed to be breakable by powerful attackers.) From disclosure attempts, I also noted there's a common misunderstanding where mail operators believe a DKIM key they do not use poses no risk. This is wrong and shows that digital signatures are sometimes difficult to get your head around. Attackers do not care if you use that key. If you want to check your own keys, you can use badkeys directly: badkeys --dkim-dns selector._domainkey.example.com [1] https://badkeys.info/ [2] https://16years.secvuln.info/ -- Hanno Böck - Independent security researcher https://itsec.hboeck.de/ https://badkeys.info/ _______________________________________________ mailop mailing list [email protected] https://list.mailop.org/listinfo/mailop
