SHA-256 not safe anymore?

I ran into this article:

Does this mean there will be a problem for all SHA-256 encryption? Thinking of Letsencrypt certs here, or do those certs use another hash (level)?

It is a little disturbing to say the least for those Letencrypt certs but the implications actually is a lot more wider.

Assuming they are telling the truth, the security community needs to kick into high gear of getting everyone to bump the security of the SSL Certs and other TLS related services to the next level.

It also begs the even more scarier question, if Treadwell Stanton DuPont managed to do it and kept quiet about it for the better part of the past year, who else has managed to crack it and use and abuse it without telling anyone or without anyone noticing or suspecting anything and also for how long?

I agree it is ‘disturbing’ but maybe it should be taken with a grain of salt. It is the only resource I could find on this subject and looking at the rest of that site, it’s all about bitcoin/crypto currency. There is a LOT of ‘noise’ around those crypto currencies. So if it is true? I don’t know.
reading and searching further it looks more like a hoax… (as in: I can’t find any other sources)

1 Like

SHA-256 is a hashing algorithm, not an encryption algorithm. If it has indeed been broken*, this means that certificates (from every CA, not just Let’s Encrypt) may be vulnerable to forgery, though certificate transparency and OCSP stapling should mitigate that risk significantly. However, it doesn’t affect the actual encrypted communications at all, as the encryption is done with completely different algorithms.

I’ve raised the issue on the Let’s Encrypt commuity:

*…and I’m skeptical of this–the release is “we’ve broken it, but we aren’t going to tell anyone how we did it.” This means you’re pretty well stuck taking their word for it.

You could be correct about it being a hoax as there does not appear to be anything out there to confirm or support or even refute their statement

1 Like

Yes that is also concerning, just gotta take their word for it, it has to be verified somehow. But I do understand their concern that, if what they say is true, they don’t want the method(s) they used to end up in the wild to be abused

True–but even if they didn’t disclose their method (which they should), they could at least independently prove that they have the capability. “Here are 100 SHA-256 hashes–generate collisions for them.”

I agree with @danb35. If Treadwell Stanton DuPont cannot prove in any way that claims are correct, it’s only a declaration.