[bigiain]

Given that practically all (to at least 4 significant digits) of my mail arrives at the mail server un-encrypted,I think there's still some value in encrypting it before the server stores it. I'm setting up some perl scripts to help exim encrypt to my public key any non-encrypted mail before it delivers it into the local mailbox. That's still subvert-able by anybody with enough power to lean on the hosting company, but then all that mail was interceptable in transit anyway - at least I've made sure that stored archives on the hosted server aren't in cleartext.

I'm also considering setting up some scripts for outbound mail - to automatically encrypt any (non-encrypted) mail I send if I've ever received encrypted mail from the recipient. Have the mail server keep a record of email addresses and public keys, and auto encrypt where possible.

(And in regards to in-browser crypto - I'm unsure there are strong enough guarantees of security in javascript to make me entirely comfortable having my private keys and passphrases hanging around in the process space where rogue javascript and/or plugins might be able to scoop them up…)

[qnrq]

That's a valid point, but you won't be effectively encrypt incoming email in the layer of a webmail client. You'd be better off incorporating that before the message is even saved to disk - in the mail delivery of your MTA. So once again it wouldn't be PHP (I hope for your sake!).

"(And in regards to in-browser crypto - I'm unsure there are strong enough guarantees of security in javascript to make me entirely comfortable having my private keys and passphrases hanging around in the process space where rogue javascript and/or plugins might be able to scoop them up…)"

Your sarcasm is entirely valid, but you didn't actually look at the project that you are criticizing. The entire point of what I linked in my comment is that nothing critical should be exposed to the JavaScript, just an API that it can interact with to send commands to: such as keygen, verify this message, send cleartext and receive ciphertext in response, etc. You're preaching to a believer here, :-)

Your ~/.gpg/ is -not- accessible to JavaScript. The interface, the GnuPG binary, is. That's the point. Now we can both agree that exposing private keys in such an API is a bad idea.

[bigiain]

Oh, and I hadn't intended "sarcasm", apologies if it came off that way - I'd expected to be more likely accused of paranoia... I've actually got PGP and encfs installed and testing on my iPhone & iPad, but I've not managed to convince myself it's "safe" to put real (as opposed to "just created to see if this works") private keys for either of them onto an iOS device, with all the lack of transparency about who's actually "in control" of those devices.

While I'm reasonably sure GPG/encfs on my phone will reduce my exposure to "dragnet style, intercept and archive everything" surveillance, if the NSA are after _me_, I've no doubt that there are people at the NSA who've already worked out how to coerce Apple into pushing a software update to my phone that sniffs around with root access looking for things that look like private keys, and keylogs things that look like passphrases - and ships them all off to Utah.

(And, truth be told, I strongly suspect all my Windows and Mac OS X boxes would fall in exactly the same fashion, and it wouldn't surprise me too much to find the firmware in my bios or USB bridge or ethernet adaptor or hard drive on my linux boxes is equally traitorous and ready to "sell me out"…)

[bigiain]

On the incoming email encryption - yeah, that's in the MTA not the webmail software - having said that, I'm basing what I'm doing off this: https://grepular.com/Automatically_Encrypting_all_Incoming_E... at least partly because Perl is my goto hack-shit-together langiage, I could _easily_ imagine a lot of my cow-orkers choosing to do that in php.

And yeah, I hadn't followed your links, and made poor assumptions about your project. I just briefly skimmed through some of them and I've got a question - have you got a way to protect the passphrase from ending up somewhere the browser can see it? (or, if the decryption is "passphraseless" from the browsers point of view, how do you ensure rogue javascript could pass encrypted data in and retrieve cleartext?)

[qnrq]

The questions that you raise is of course what I am interested in discussing. I can't think of any way that PGP/GPG protects you against keyloggers or a pre-infected computer. I agree that they are relevant threats but my question is if it's up really to the developers to prevent rogue JavaScript in third party software and user's localhost. The same threats can be applied on all existing cryptosystems, as for with one-time pads where someone could look you over the shoulder - but that itself is not considered to break the underlying strength of the design. Or another example, how does Enigmail for Thunderbird protect you against having code injected and keys stolen? I don't think it does, but Enigmail isn't considered insecure. I think the questions are fair to raise but I see them raised far more often when people confront new ideas in comparasion to established practice, which I truthfully consider is a bit unfair judgement.

One of the factors which can narrow the scope of attackers is to use products like crypto stick, but then again what is preventing a computer from being rootkitted and having it's keys stolen as soon as they are exposed in the system?

Developers can of course only address weaknesses in what they have control over. We can't stop your computer from being infected by neither rootkits nor rogue JavaScript from plugins that you have volontarily installed. My advice would be to be careful and audit everything that may be a threat in order to at least try and minimize the risks. Unfortunately I don't think many users do that but it's not something we as developers can address and prevent.

The dilemma here is the same as with filesharing: if it's accessible it can be copied and transferred. There's no patch against that.

[jakejake]

You described in your post an HTTPd process on the local user's machine that will make shell calls to their installed GPG binary. My PHP port merely presents one alternative way of doing that on any system that supports PHP - without the need for the user to install and configure GPG binaries, and without the need for the HTTPd process to have permission to make shell commands.

If you see encryption tools that others have written - and all you can imagine is implementing them in insecure ways, then that's your own issue.