Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by sys42590 on Thursday January 19, 2023 @04:10PM
(#34445651)
ZeroSSL left an uncanny impression on me when for some reason acme.sh developers made them default instead of Let's Encrypt. This prompted me to switch to a different client (just in case of further worsening of Let's Encrypt support by acme.sh).
[ Parent
| on HN ]
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by leetnewb on Thursday January 19, 2023 @04:21PM
(#34445769)
Which client did you end up on? The list is somewhat overwhelming.
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by sys42590 on Thursday January 19, 2023 @04:23PM
(#34445800)
dehydrated, as it has little dependencies.
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by Ennea on Thursday January 19, 2023 @04:59PM
(#34446258)
Going to throw another hat into the ring here: I use acme-tiny [1], which is a single file ACME client written in Python in under 200 lines. The idea behind it is that you can fully read and understand everything it does without spending too much time on it. I really like this approach, so I went ahead and started using it, and have been for a few years now.
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by seized on Thursday January 19, 2023 @08:36PM
(#34448119)
I too am moving away from acme.sh for the same reason. Dehydrated looks nice but I started using goacme.
https://github.com/go-acme/lego
I wasn't set on only bash though.
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by netheril96 on Thursday January 19, 2023 @10:33PM
(#34448878)
I prefer ZeroSSL to Let's Encrypt. ZeroSSL has no rate limit, and most importantly they have full ECC support. With Let's Encrypt, even if I request for an ECC cert, the intermediate CA is still RSA, drastically increasing the certificate size (they have their reasons of compatibility, but I don't care about that).
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by Ayesh on Friday January 20, 2023 @03:24AM
(#34450415)
LetsEncrypt now has an ECC root and intermediates. You have to request the account ID to be included, and after which, the intermediate and root certificates will be ECC. More information here: https://community.letsencrypt.org/t/ecdsa-availability-in-pr...
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by netheril96 on Friday January 20, 2023 @07:28AM
(#34452176)
The alternative you suggest has a longer chain of certificates, and more difficult setup. Using ZeroSSL is way easier with less bytes on the TLS handshake.
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by Ayesh on Friday January 20, 2023 @07:38AM
(#34452277)
Do you have a test host with the Zerossl chain that you speak of? Use https://aye.sh if you want to try a host using the ECC chain from LE.
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by netheril96 on Friday January 20, 2023 @08:55AM
(#34453149)
So the article is outdated I guess. The length of the chain is the same now.
I'll consider switching back to Let's Encrypt once this setup doesn't require a whitelist.
Re:ZeroSSL: XSS to session hijacking, stealing a private key...
by dilyevsky on Thursday January 19, 2023 @11:00PM
(#34449074)
I believe zerossl chain (really sectigo) is trusted by more devices than the new isrg root (mostly old unupdated ones). Also zerossl has fewer limits in their acme implementation. Downsides are zerossl has some questionable security practices and also I think zerossl either dont support tls-alpn-01 validation or it’s just broken