We should talk about Werner Koch's response gpg.fail on the oss-security mailing list.

openwall.com/lists/oss-securit…

Yes, and actually the only serious bug from their list.


Koch either didn't watch the talk, he is in such defense of his own ego that he can't see how serious the bugs were, or he's tacitly admitting that PGP is not a serious recommendation.

Can you distinguish between these three explanations?

Could it be all of them are true?

Impact

While this may allow remote code execution (RCE), it definitively causes memory corruption.


Good research.


I think this sarcastic quip is what reveals Werner Koch's opinion about the security researchers and their work.

The rest of his email is measured (and partly responding to other mailing list participants rather than the disclosure directly).

Als Antwort auf Soatok Dreamseeker

I think 2026 should be the year that we make PGP irrelevant.

Not just GnuPG (Koch's implementation), but the entire OpenPGP ecosystem.

Most cryptographers I talk to gave up on PGP over a decade ago.

(After seeing the arrogance and dismissiveness that bled through Koch's oss-security email, who can blame them?)

If you're a country whose government mandates the use of PGP, even in obscure places, let's talk about how to replace PGP.

Als Antwort auf Soatok Dreamseeker

@Soatok Dreamseeker so, are you suggesting we switch to alternative openpgp implementations, such as sequoia or a different email encryption standard, such as S/MIME? I've refrained from using the latter in the past because it is not a web-of-trust solution, but requires trusted third parties, which generally means paying some blessed organization. I'd consider using it if there were non-commercial signing entities, like letsencryt does for TLS. Till now I've rather supported efforts of replacing gpg with sq, for example helping to package that for alpine.
Als Antwort auf Soatok Dreamseeker

@elrido soatok.blog/2024/11/15/what-to…
Als Antwort auf Soatok Dreamseeker

Als Antwort auf Soatok Dreamseeker

I'm all for replacing outdated technologies, but I don't think getting rid of OpenPGP entirely could happen any time soon. especially RPM based Linux distros heavily rely on it for package signing ... (but at least things are migrating away from using GnuPG for this purpose)
Als Antwort auf Soatok Dreamseeker

@pemensik you cant do that. Email is a de facto standard that works *everywhere*. You cant ditch that for another protocol before decades, and thats supposing tge new stuff has all the required features

It's good to be idealist but the industrial world needs practical solutions. I would love a good replacement for PGP, but lets be honest, there is no tool that can protect files and email before transfer NOW.

Maybe the path forward is fixing pgp, not yeeting it.

Als Antwort auf Soatok Dreamseeker

@f4grx @pemensik You cannot fix emails. I'm sorry, but that's just not going to happen.

You either replace email or you give up entirely. There's too much political power in the hands of people that do not want anything solved.

PGP is downstream of that mess. People have tried for years to fix PGP. That's what Sequoia allegedly was going to be.

That still hasn't happened, because PGP implementations refuse to say "we only support the newest stuff, and only the features and use cases that need to exist".

Why was the \f flag even in GnuPG? That shouldn't be a feature. If you designed a CTF challenge with that feature the judges would tell you it's too obvious and stupid, and to go back to the drawing board.

Als Antwort auf Soatok Dreamseeker

@Soatok Dreamseeker As some one familiar with the whole stack when it comes to email, it is not that there is any political power holding back adding in something to replace PGP with a working RFC, it is that there is a stupidly large number of projects involved to make it all work and it requires getting every one on board. It is why new RFCs tend to get adopted slowly for it. Takes it's way to make it through the entire ecosystem.

If you want a closer to home example, look at the encryption stuff you've mentioned working on for ActivityPub/Mastodon. So once you get it implemented for Mastodon, there is still every other ActivityPub server that also needs to support it as well.

And here we are talking about a ecosystem that is incredibly small compared to the entire stack that is email.

Als Antwort auf Soatok Dreamseeker

wouldn't it be enough to propose RFC for secure email handling by email clients? Creating completely new protocol might hit problems solved already in existing protocols. Or sharing the same problems the old ones failed to solve well enough. For example user friendly, decentralized key distribution has no great solution, if it should be secure at the same time?
Als Antwort auf Petr Menšík

@pemensik

wouldn't it be enough to propose RFC for secure email handling by email clients?


No. The problems with encrypted email are unsolvable by RFCs.

Creating completely new protocol might hit problems solved already in existing protocols.


This is why github.com/soatok/awoo-specifi… is still an empty repo: It's work I planned to do after I solved key transparency for Fedi and E2EE happens.

For example user friendly, decentralized key distribution has no great solution, if it should be secure at the same time?


Literally, soatok.blog/2025/12/15/announc…

Als Antwort auf Orman

@orman @pemensik And there's half a century of legacy baked into the email RFCs and software that implements them.

Most end users don't care about RFCs or protocols. They care about their user experience.

Thus, I'm going to write an encrypted alternative to email, rather than try to fix email itself.

It will be encrypted always. There is no plaintext mode. The entire failure mode of "oops I replied to an encrypted email with plaintext" is a stupid problem to have.

Dieser Beitrag wurde bearbeitet. (Thursday, January 1, 2026, 21:16)
Als Antwort auf Soatok Dreamseeker

@Soatok Dreamseeker If you don't mind me saying, you are looking are likely looking at creating a encrypted messaging system than a actual alternative to email.

Saying you are creating a alternative to email means tackling a whole hell of a lot of infrastructure that needs replacing.

I'm not saying you should not start work on something like that, I just think you massively underestimate how much is involved or some of the unique issues bodies the server can't parse as they are encrypted bring(such as it makes adversarial handling much harder as writing a replacement for Spamassassin needs to use headers only then, meaning you now need to split it in two as you need part to run on the server and part to run on the viewing client). Writing a replacement for Sieve means similar issues as well.

Als Antwort auf Soatok Dreamseeker

Als Antwort auf Soatok Dreamseeker

It’s amazing to me how many people think email is secure, with or without the S/Mime extension. Even people who should know better (HR, finance, etc).

I continually have to tell people that email is like a postcard sent through postal service. Anyone who has access to it can read it. “Would you feel comfortable sending a postcard with your credit card information plainly written on it?”

Als Antwort auf FurryBeta

@FurryBeta oh, that's nothing. I have yelled repeatedly because email CAN be secure. We fucking have TLS1.3 transport. We've had good local filtering mechanisms since the 1990's.

And guess what? The big three are the biggest fucking impediment by refusing to enforce TLS, intentionally sabotaging known secure methods, and continually misleading users and customers.

Als Antwort auf RootWyrm 🇺🇦

@rootwyrm @FurryBeta TLS 1.3 gives transport-layer encryption.

The fact that TLS isn't already universally a thing for email is annoying.

But the thing we're talking about (and why PGP keeps coming up) is folks want their emails confidential from their provider.

And, sorry, that's just not doable. Experts routinely reply to PGP-encrypted emails with plaintext, often quote-replying the email that was encrypted.