[Petition][Poll] We want Https on XDA!
Three days ago I had a conversation with @benkxda. During that conversation, he pointed out that XDA uses no https encryption. Therefore everybody can read our passwords and PMs when we send them to the server. What if someone replaced our downloadable files with viruses?
Is that really what we want? Neither @benkxda nor I think so. Hence we decided to create this thread.
Now you might ask what you can do to get https on our forum. The first and easiest option is probably the most effective. Vote in the poll at the top of the page.
If you've got some spare time, you can also write a short (or long) post explaining your opinion.
If we get enough votes for this petition, the admins might consider supporting the https protocol.
To ensure that everybody sees this, we want to get this on the portal. Please help us by either clicking this link or by clicking the "Tip us?" button in the right upper corner of this post.
It would also help to spread the word if you put a link to this thread into your signature.
Thanks for reading.
Announcements
4th March 2014: 1000 supporters.
2nd January 2014: bitpushr implemented https for the login form! Thanks to all supporters.
23rd December 2013: And again, doubled. 800 votes now.
1st November 2013: Another announcement by bitpushr: They "have gotten [their] CDN provider to add SSL." Moreover, he will "add this to the forums".
28th September 2013: Doubled, again. 400 now.
31st August 2013: We just hit the 200 voters mark. Thanks.
13th August 2013: We reached 100 supporters. Keep voting.
7th August 2013: bitpushr announced that the admin team is working on https. I want to say thank you to all who have voted yet. But remember, we don't have https yet. So continue to vote.
29th July 2013: This petition was created.
Click to expand...
Click to collapse
Code for the signature
Code:
[SIZE="5"[B][/B]][[B][/B]URL="http://forum.xda-developers.com/showthread.php?t=2383868"][[B][/B]COLOR="Blue"]Vote for a secure XDA: [/[B][/B]COLOR][Petition][Poll] We want Https on XDA![/UR[B][/B]L][/SI[B][/B]ZE]
Well, XDA folks, you have to take the poll serious. In days where secret services all over the world spy almost everything, the poll has two options, a secret service version as well as a normal version :angel:
But to be honest, we are not safe from those spies. Encryption can help much - not only against those spy experts, but also against the administrators in a network, for eg in the company.
Currently, we have no secured connection like SSL/TLS secured HTTPS. Login data can be stolen, every communication is held open. We need a secure connection for the whole XDA website, including linked in scripts and images and not limited to the login sequence. This is state of the art even at Google or Facebook.
benkxda said:
Well, XDA folks, you have to take the poll serious. In days where secret services all over the world spy almost everything, the poll has two options, a secret service version as well as a normal version :angel:
But to be honest, we are not safe from those spies. Encryption can help much - not only against those spy experts, but also against the administrators in a network, for eg in the company.
Currently, we have no secured connection like SSL/TLS secured HTTPS. Login data can be stolen, every communication is held open. We need a secure connection, which is state of the art at Google or Facebook.
Click to expand...
Click to collapse
All sites these days should be https. Also I want to add that it is important that https is not only added to the login itself but the entire site. To cut cost, lots of sites use http to https redirect for login only and then swtich the user back to http. Problems with that are tools for cookie hijacking, session hijacking, and tools like sslstrip. The vote should be for SITE WIDE https.
Let's face facts people. On XDA, we download things and flash to our phones, tablets or other devices. If our account is hijacked )which is so easy its not funny) then someone else can replace our material with ones that have back doors/trojans and update the posted MD5. No one would know. security is a concern for me at least.
calisro said:
All sites these days should be https. Also I want to add that it is important that https is not only added to the login itself but the entire site. To cut cost, lots of sites use http to https redirect for login only and then swtich the user back to http. Problems with that are tools for cookie hijacking, session hijacking, and tools like sslstrip. The vote should be for SITE WIDE https.
Let's face facts people. On XDA, we download things and flash to our phones, tablets or other devices. If our account is hijacked )which is so easy its not funny) then someone else can replace our material with ones that have back doors/trojans and update the posted MD5. No one would know. security is a concern for me at least.
Click to expand...
Click to collapse
True, only full secured websites are really secured. Thanks for this hint, will edit my prior post.
calisro said:
All sites these days should be https. Also I want to add that it is important that https is not only added to the login itself but the entire site. To cut cost, lots of sites use http to https redirect for login only and then swtich the user back to http. Problems with that are tools for cookie hijacking, session hijacking, and tools like sslstrip. The vote should be for SITE WIDE https.
Let's face facts people. On XDA, we download things and flash to our phones, tablets or other devices. If our account is hijacked )which is so easy its not funny) then someone else can replace our material with ones that have back doors/trojans and update the posted MD5. No one would know. security is a concern for me at least.
Click to expand...
Click to collapse
Of course, it should be added to the entire site. However, I didn't even think about the downloading thing. That's definetly true and I'll add that.
Feel free to spread the word.
Thank you very much. :good:
benkxda said:
True, only full secured websites are really secured. Thanks for this hint, will edit my prior post.
Click to expand...
Click to collapse
Posted at the same time. :laugh:
benkxda said:
True, only full secured websites are really secured. Thanks for this hint, will edit my prior post.
Click to expand...
Click to collapse
Not fully correct.
NSA is getting also access to https secured connections.
http://www.dailytech.com/FBI+NSA+Wa...Keys+from+Internet+Companies/article32046.htm
Mardon said:
Not fully correct.
NSA is getting also access to https secured connections.
http://www.dailytech.com/FBI+NSA+Wa...Keys+from+Internet+Companies/article32046.htm
Click to expand...
Click to collapse
That's right, but our main concern should be the (bad) hackers. It is difficult to stop the NSA, you know.
Mardon said:
Not fully correct.
NSA is getting also access to https secured connections.
http://www.dailytech.com/FBI+NSA+Wa...Keys+from+Internet+Companies/article32046.htm
Click to expand...
Click to collapse
This must be verified first, but frankly I really believe, they try to get those master keys. But they would need a master key to get access. At least, an encryption keeps out most assailants.
nikwen said:
That's right, but our main concern should be the (bad) hackers. It is difficult to stop the NSA, you know.
Click to expand...
Click to collapse
Right https is much better i agree
If NSA or FBI or who else gets the masterkeys there also exist a chance for others (hackers) to get the keys too.
I think the whole internet needs a new full encrypted security protocol in future where the keys are randomly changed and such things like masterkeys only working a few hours to minimize the hacking risks.
But thats offtopic i think
Mardon said:
Right https is much better i agree
If NSA or FBI or who else gets the masterkeys there also exist a chance for others (hackers) to get the keys too.
I think the whole internet needs a new full encrypted security protocol in future where the keys are randomly changed and such things like masterkeys only working a few hours to minimize the hacking risks.
But thats offtopic i think
Click to expand...
Click to collapse
Oh yes, indeed I recently thought almost the same. And maybe we are a bit special picky, hope the "normal" users can keep up that indignation or sometimes outrage on these spy stuff. Also true, off topic.
Mardon said:
Right https is much better i agree
If NSA or FBI or who else gets the masterkeys there also exist a chance for others (hackers) to get the keys too.
I think the whole internet needs a new full encrypted security protocol in future where the keys are randomly changed and such things like masterkeys only working a few hours to minimize the hacking risks.
But thats offtopic i think
Click to expand...
Click to collapse
you realize there aren't one set of master keys for all certificates right? lol. Each certificate has a master key owned by the company owning the cert. If facebook gives them their master keys that doesn't mean they can snoop your xda or bank account traffic.
ok back on topic! I digress!
Mardon said:
Not fully correct.
NSA is getting also access to https secured connections.
http://www.dailytech.com/FBI+NSA+Wa...Keys+from+Internet+Companies/article32046.htm
Click to expand...
Click to collapse
Just saying, but on HTTPS stuff that we use, we use forward-secret HTTPS. Meaning the "private key" for the site is of no use for decrypting past connections. That's becoming more popular for larger sites these days, but I started looking into it a while ago, and it is ready to use now. Look for a key exchange method of DHE or ECDHE
As such, the only value in obtaining such keys would be to spoof future connections. If someone is that determined to target YOU individually with spoofed or MITM'd connections, you should be worrying about other things (it would be fairly impractical to mount a widescale meaningful attack).
If you are concerned, you should look into the issues with the CA system who issue SSL keys - an SSL certificate can be signed by ANY of them, and there's a number of CAs who are somewhat sketchy in trust... Tl;dr if an active attacker wants a key for your site to spoof it, he can get it. It won't be the same one (cannot decrypt legit traffic), but can be used to impersonate the site.
pulser_g2 said:
Just saying, but on HTTPS stuff that we use, we use forward-secret HTTPS. Meaning the "private key" for the site is of no use for decrypting past connections. That's becoming more popular for larger sites these days, but I started looking into it a while ago, and it is ready to use now. Look for a key exchange method of DHE or ECDHE
As such, the only value in obtaining such keys would be to spoof future connections. If someone is that determined to target YOU individually with spoofed or MITM'd connections, you should be worrying about other things (it would be fairly impractical to mount a widescale meaningful attack).
If you are concerned, you should look into the issues with the CA system who issue SSL keys - an SSL certificate can be signed by ANY of them, and there's a number of CAs who are somewhat sketchy in trust... Tl;dr if an active attacker wants a key for your site to spoof it, he can get it. It won't be the same one (cannot decrypt legit traffic), but can be used to impersonate the site.
Click to expand...
Click to collapse
Thanks for the info. I didn't know that.
Not a techie nor from a part of the world affected by PRISM (?) but still having read all this I'm inclined to say i second this motion
nikufellow said:
Not a techie nor from a part of the world affected by PRISM (?) but still having read all this I'm inclined to say i second this motion
Click to expand...
Click to collapse
Great.
Are you sure that you are not affected? Everyone is, some more, some less.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
(http://upload.wikimedia.org/wikipedia/commons/5/5c/Boundless-heatmap-large-001.jpg)
We've reached over 50 votes.
nikwen said:
We've reached over 50 votes.
Click to expand...
Click to collapse
Nice. So, some people know about HTTPS and encryption and prefer that. Not only because of the secret services of the "Five Eyes", but also to protect me from curious network administrators. There are surely more on XDA, who want to support this necessary petition.
EDIT: Oh, please don't misunderstand, I did not want to protect the secret services from other countries, as they might be not better in privacy protection, for eg the German secret service called "BND" seems to be the sixth eye. Again, I did not want to say only those five do bad things.
As lots of users don't know / care about encryption, a secured https connection with XDA might sensibilize at least some.
So, I support your request.
rog_star said:
As lots of users don't know / care about encryption, a secured https connection with XDA might sensibilize at least some.
So, I support your request.
Click to expand...
Click to collapse
Yeah, I hope so.
Thanks for voting.
Some of us use Textsecure as replacement for Stock SMS app. Textsecure provides encryption for your SMS. However, my recommendation is: stay away or at least don't update to 2.X... versions.
The developer has introduced Google Cloud Messaging, which means that even if your sms are secure, the fact you are using the app will be recorded in Google Centralized database. In addition, he removed the ability of the user to regenerate new identity key. In last couple of releases, he forced the user to allow the app to contact the internet (otherwise, the app would crash). That is even if you compile the app from sources, which I did a couple of hours ago. If you download the app from Store, you can't even use it without Google account and GSF, the latter will record your every keystroke including the password used to encrypt the messages. In further addition, the app is only available through Googleplay and the developer is actively resisting third party distribution. If that is not enough, you should know that Whisper systems is owned by Twitter, which is a red flag in of itself. The code is growing larger and is more difficult to examine for back door purposes.
My advice: stay away from this development, which in my view is compromised...
Edit. In January of this year, the developer left Twitter. Interestingly, he is still working on Textsecure and it is published under Whisper, which is Twitter. About the same time, all those things described above started to happen. Also interesting is that the developer was put on federal watch list and was continuously harrased by various agencies when flying. So, I wouldn't be surprised to learn that his new employer is the previous harraser...
All more reasons to stay away from this app.
optimumpro said:
Some of us use Textsecure as replacement for Stock SMS app. Textsecure provides encryption for your SMS. However, my recommendation is: stay away or at least don't update to 2.X... versions.
The developer has introduced Google Cloud Messaging, which means that even if your sms are secure, the fact you are using the app will be recorded in Google Centralized database. In addition, he removed the ability of the user to regenerate new identity key. In last couple of releases, he forced the user to allow the app to contact the internet (otherwise, the app would crash). That is even if you compile the app from sources, which I did a couple of hours ago. If you download the app from Store, you can't even use it without Google account and GSF, the latter will record your every keystroke including the password used to encrypt the messages. In further addition, the app is only available through Googleplay and the developer is actively resisting third party distribution. If that is not enough, you should know that Whisper systems is owned by Twitter, which is a red flag in of itself. The code is growing larger and is more difficult to examine for back door purposes.
My advice: stay away from this development, which in my view is compromised...
Edit. In January of this year, the developer left Twitter. Interestingly, he is still working on Textsecure and it is published under Whisper, which is Twitter. About the same time, all those things described above started to happen. Also interesting is that the developer was put on federal watch list and was continuously harrased by various agencies when flying. So, I wouldn't be surprised to learn that his new employer is the previous harraser...
All more reasons to stay away from this app.
Click to expand...
Click to collapse
And here is some more fresh evidence. Today I posted this info on Cyanogen site related to Textsecure Push for CM.
http://www.cyanogenmod.org/blog/whisperpush-secure-messaging-integration
The site says it is neither censored no monitored. Within 5 minutes, the post has disappeared... . So, stay away from this app as the development has been compromised. In my view, of course...
You have no clue what youre talking about.
Corndude said:
You have no clue what youre talking about.
Click to expand...
Click to collapse
Thanks, pal... for a very, very thorough, thoughtful and factual argument.
Edit: by the way, what does no gapps project have to do with textsecure being compromised?
Thanks for the heads up. Something is really amiss, and I won't want to directly experience it. I'm staying away from TextSecure for sure.
abdelazeez said:
Thanks for the heads up. Something is really amiss, and I won't want to directly experience it. I'm staying away from TextSecure for sure.
Click to expand...
Click to collapse
Most messenger apps today work with Google Push Notifications, seems to be no problem for people there. Funny that it is here. As for SMS, I would never use that through another app. Besides, the phone carrier companies save those probably too, whats so different with that you said ? Text Secure is a very nice app I think. Right now people on iOS don't have that app yet, which makes it hard to establish in mixed system userbases among people. But I hope that will change.
Besides, most people here probably use Twitter. Funny to complain about something that might be related to Twitter then, isn't it ?
Wolfseye
wpkwolfseye said:
Most messenger apps today work with Google Push Notifications, seems to be no problem for people there. Funny that it is here. As for SMS, I would never use that through another app. Besides, the phone carrier companies save those probably too, whats so different with that you said ? Text Secure is a very nice app I think. Right now people on iOS don't have that app yet, which makes it hard to establish in mixed system userbases among people. But I hope that will change.
Besides, most people here probably use Twitter. Funny to complain about something that might be related to Twitter then, isn't it ?
Wolfseye
Click to expand...
Click to collapse
The difference is that Textsecure/Whisperpush/CMpush tell you your SMS are encrypted. If they are indeed encrypted and there are no backdoors, your carrier (and others) can only get encrypted SMS (good luck to them trying to decipher). All other SMS apps are in plain text. In my view earlier versions of Textsecure are indeed secure. Starting from version 2.X, we no longer know that considering all the facts I mentioned in the OP.
You should really get your facts straight. Twitter bought Whisper Systems in 2011, mainly to get Moxie and the other Whisper Systems folks to work for them.
Moxie went on to lead Twitters security team. Twitter allowed them a month or so after they aquired Whisper Systems to open source their apps TextSecure and RedPhone. In January 2013 Moxie left Twitter and started Open Whisper Systems with a few others. They took the newly open sourced apps and developed them further.
This is also covered in their FAQ.
You can see all of their code on GitHub.
And if you don't have GAPPS installed, you will simply get a message that you won't be able to use push messages and that's it. Several friends of mine use it for SMS only, with Xprivacy restricting the internet access. It doesn't crash or anything.
If you experience this, you may either have a problem with your build or it's a bug specific to your device/Android version.
Moxie also wrote exactly why he doesn't want TextSecure to be released via F-Droid: for security reasons. They use central signing, which may very well compromise the update channel.
The whole discussion can be found in the most infamous thread in their GitHub: #127
lindworm said:
You should really get your facts straight. Twitter bought Whisper Systems in 2011, mainly to get Moxie and the other Whisper Systems folks to work for them.
Moxie went on to lead Twitters security team. Twitter allowed them a month or so after they aquired Whisper Systems to open source their apps TextSecure and RedPhone. In January 2013 Moxie left Twitter and started Open Whisper Systems with a few others. They took the newly open sourced apps and developed them further.
This is also covered ir FAQ.
You can see all of their code on GitHub.
And if you don't have GAPPS installed, you will simply get a message that you won't be able to use push messages and that's it. Several friends of mine use it for SMS only, with Xprivacy restricting the internet access. It doesn't crash or anything.
If you experience this, you may either have a problem with your build or it's a bug specific to your device/Android version.
Moxie also wrote exactly why he doesn't want TextSecure to be released via F-Droid: for security reasons. They use central signing, which may very well compromise the update channel.
The whole discussion can be found in the most infamous thread in their GitHub: #127
Click to expand...
Click to collapse
Which fact did I not get straight? You can't get the app anywhere other than from Googleplay and for Googleplay you need GSF, which records your every keystroke. And by the way, try to restrict getnetworkinfo in internet settings in Xprivacy and the app will crash as soon as you try to open a conversation (checked on several devices). And why was it necessary to prevent users from generating new identity key? Why not have an app available on Whisper's github, as many devs do. And by the way, I asked the same questions on github and f-droid threads and in response got a suggestion to build an equivalent of Google's GCM, so then Moxie would stop using Google.
optimumpro said:
Which fact did I not get straight? You can't get the app anywhere other than from Googleplay and for Googleplay you need GSF, which records your every keystroke. And by the way, try to restrict getnetworkinfo in internet settings in Xprivacy and the app will crash as soon as you try to open a conversation (checked on several devices). And why was it necessary to prevent users from generating new identity key? Why not have an app available on Whisper's github, as many devs do. And by the way, I asked the same questions on github and f-droid threads and in response got a suggestion to build an equivalent of Google's GCM, so then Moxie would stop using Google.
Click to expand...
Click to collapse
You are not even trying to learn/understand why things are done the way they are done, but instead chose to blast an open source project by a security expert who has spoken at defcon various times and who is on a national security list and gets severely hassled by the TSA every time he tries to travel because of his involvement with secure communication projects.
You don't show the slightest form of objectiveness either. The truth content of what you are writing varies between "flat out wrong" and "there is a reason for how they do it that way, which you either didn't care to research or willingly ignored".
1. You can sideload the apk either from http://apps.evozi.com/apk-downloader/ or any of the dozens of sites that mirror packages from the app store.
They do not provide apks because it is a security risk: there is no automated upgrade channel from where a user can get a new version which may fix serious security flaws.
Everybody who is able to compile from source however should understand the importance of updating regularly and can do so on his/her own.
Moxie stated all of that in the github ticket I linked to.
2. GSF doesn't record your keystrokes.
3. If you had bothered to look it up, getNetworkInfo returns if a certain interface (like wifi) is used for internet.
This leaks no interesting information whatsoever. And it especially doesn't mean that TextSecure doesn't work without internet, because this permission does not give an app internet access. Xprivacy actually expects this behaviour by apps, that's why those fields are by default not restricted even if you restrict internet access of an app.
The program crashes without this, because it expects to get a needed value returned, which you chose to block. This is not something they willingly built in, to stop you from using it without Google Play.
If you can't manage the complexity of the permissions, you should use a simple firewall like AFwall+ to restrict internet access.
4. This was probably removed because it doesn't add any significant security and adds clutter to the user interface, because average users have no idea what it's for. The identity keys you are talking about are long term identity keys. TextSecure uses different keys in every message and actually uses the most secure protocol I know of. It has excellent forward secrecy, future secrecy and deniability. More so than OTR, which it is derived from.
You can learn more about that in their blog:
https://whispersystems.org/blog/simplifying-otr-deniability/
https://whispersystems.org/blog/asynchronous-security/
https://whispersystems.org/blog/advanced-ratcheting/
5. You asked them to not use the only free world wide push network that has contracts with all major providers to not kill idle TCP connections.
Moxie always answered that they would love to use something else, but none exists. And that they don't have the resources to build a push network themselves.
This is all in the comments to https://whispersystems.org/blog/the-new-textsecure/ and on ycombinator:
https://pay.reddit.com/r/Android/co..._cyanogenmod_is_integrating/cdyfxhm?context=3
https://pay.reddit.com/r/Android/co..._cyanogenmod_is_integrating/cdyfrv0?context=3
They are however working on using emails as identifiers and websockets as an alternative to GCM. Websockets are already implemented on the server side and people are working on the client side.
Right now you can use encrypted SMS without GCM, no problem at all. If you want to use it over the internet, you can help to speed up the websocket development:
https://github.com/WhisperSystems/TextSecure/issues/1000
lindworm said:
You are not even trying to learn/understand why things are done the way they are done, but instead chose to blast an open source project by a security expert who has spoken at defcon various times and who is on a national security list and gets severely hassled by the TSA every time he tries to travel because of his involvement with secure communication projects.
You don't show the slightest form of objectiveness either. The truth content of what you are writing varies between "flat out wrong" and "there is a reason for how they do it that way, which you either didn't care to research or willingly ignored".
1. You can sideload the apk either from http://apps.evozi.com/apk-downloader/ or any of the dozens of sites that mirror packages from the app store.
They do not provide apks because it is a security risk: there is no automated upgrade channel from where a user can get a new version which may fix serious security flaws.
Everybody who is able to compile from source however should understand the importance of updating regularly and can do so on his/her own.
Moxie stated all of that in the github ticket I linked to.
2. GSF doesn't record your keystrokes.
3. If you had bothered to look it up, getNetworkInfo returns if a certain interface (like wifi) is used for internet.
This leaks no interesting information whatsoever. And it especially doesn't mean that TextSecure doesn't work without internet, because this permission does not give an app internet access. Xprivacy actually expects this behaviour by apps, that's why those fields are by default not restricted even if you restrict internet access of an app.
The program crashes without this, because it expects to get a needed value returned, which you chose to block. This is not something they willingly built in, to stop you from using it without Google Play.
If you can't manage the complexity of the permissions, you should use a simple firewall like AFwall+ to restrict internet access.
4. This was probably removed because it doesn't add any significant security and adds clutter to the user interface, because average users have no idea what it's for. The identity keys you are talking about are long term identity keys. TextSecure uses different keys in every message and actually uses the most secure protocol I know of. It has excellent forward secrecy, future secrecy and deniability. More so than OTR, which it is derived from.
You can learn more about that in their blog:
https://whispersystems.org/blog/simplifying-otr-deniability/
https://whispersystems.org/blog/asynchronous-security/
https://whispersystems.org/blog/advanced-ratcheting/
5. You asked them to not use the only free world wide push network that has contracts with all major providers to not kill idle TCP connections.
Moxie always answered that they would love to use something else, but none exists. And that they don't have the resources to build a push network themselves.
This is all in the comments to https://whispersystems.org/blog/the-new-textsecure/ and on ycombinator:
https://pay.reddit.com/r/Android/co..._cyanogenmod_is_integrating/cdyfxhm?context=3
https://pay.reddit.com/r/Android/co..._cyanogenmod_is_integrating/cdyfrv0?context=3
They are however working on using emails as identifiers and websockets as an alternative to GCM. Websockets are already implemented on the server side and people are working on the client side.
Right now you can use encrypted SMS without GCM, no problem at all. If you want to use it over the internet, you can help to speed up the websocket development:
https://github.com/WhisperSystems/TextSecure/issues/1000
Click to expand...
Click to collapse
Your original statement was that I got my facts wrong. Since you have not cited any instance where I came up with a wrong fact, I will address your opinions.
Number one: you say GSF does not record keystrokes. How do you know? Have you seen the source (which is closed)? If you did, you work for Google and then everything you say is propaganda that has zero factual value. If you don't, then you are just speculating. You pick whichever is worse. If you use Google proprietary blobs, your device is totally open and there is no security measure/app on earth that is effective against this. That GSF phones home at regular intervals and transmits data there is a known fact. You can use encryption from Mars and yet it won't work because raw data (before encryption) is open to Google. As another user noted, having GSF and other closed source apps is like having a lock installed on your house door and not knowing who has access to it besides you.
Number two: inability to generate new identity key: It was there for a reason, the same way PGP or GPG keys have the ability to be limited in time, revoked or regenerated. It is a good security standard and removing it represents weakening. Clutter? LOL. A regular user wouldn't even be able to find it. Certainly, it does not pop up anywhere, one has to find it.
Number three: Sideload or compiling: a regular user will do neither, he/she will simply download the app from the market, which means he has to have Google blobs. Or you are suggesting that users should download the app from the market and then remove GSF and other Googleapps? LOL again.
As I said earlier, Moxie's argument that allowing third party apps on your device is a greater security risk than having closed source blobs is wrong and grand BS (especially coming from someone who is considered a security expert). It is security through obscurity, which is no security at all. The value of his open source project is completely defeated by having closed source blobs by a known private branch of known three letter agencies.
Now, these are facts. Let's get to opinions. I think that this deliberate weakening of security (again coming from a security expert) is a strong indication that development and/or developer has been compromised. And that is why I recommend to stay away from this app. But that is just my opinion, which is nonetheless based on facts.
optimumpro said:
Your original statement was that I got my facts wrong. Since you have not cited any instance where I came up with a wrong fact, I will address your opinions.
Click to expand...
Click to collapse
Do you even read what I write?
If that is not enough, you should know that Whisper systems is owned by Twitter, which is a red flag in of itself.
Click to expand...
Click to collapse
As I explained he does now work there any more.
You seem to have noticed that too:
Edit. In January of this year, the developer left Twitter. Interestingly, he is still working on Textsecure and it is published under Whisper, which is Twitter.
Click to expand...
Click to collapse
Are you kidding me? How the flying **** did you get to this conclusion? The company that was bought by twitter was Whisper Systems.
They are publishing the new source under Open Whisper Systems. (none of those was ever called Whisper)
See the difference? They also state this here: http://support.whispersystems.org/customer/portal/articles/1474591-is-textsecure-owned-by-twitter-
And here is some more fresh evidence. Today I posted this info on Cyanogen site related to Textsecure Push for CM.
http://www.cyanogenmod.org/blog/whis...ng-integration
The site says it is neither censored no monitored. Within 5 minutes, the post has disappeared... . So, stay away from this app as the development has been compromised. In my view, of course...
Click to expand...
Click to collapse
So you are saying CyanogenMod is part of this grand conspiracy of yours? Come on...
GSF, which records your every keystroke.
Click to expand...
Click to collapse
Number one: you say GSF does not record keystrokes. How do you know? Have you seen the source (which is closed)? If you did, you work for Google and then everything you say is propaganda that has zero factual value. If you don't, then you are just speculating. You pick whichever is worse. If you use Google proprietary blobs, your device is totally open and there is no security measure/app on earth that is effective against this. That GSF phones home at regular intervals and transmits data there is a known fact. You can use encryption from Mars and yet it won't work because raw data (before encryption) is open to Google. As another user noted, having GSF and other closed source apps is like having a lock installed on your house door and not knowing who has access to it besides you.
Click to expand...
Click to collapse
It's a binary blob and it sends data to google, but you have no proof whatsoever if it records keystrokes. You can know if you want to tough. Decompile it and analyze it. I don't like binary blobs, but you can't just say they do something without having any proof. I may not be able to guarantee that they don't do something, because I have not personally decompiled and analyzed every bit of it, but until you have and have proof that it does do something you can't just claim it does.
Number two: inability to generate new identity key: It was there for a reason, the same way PGP or GPG keys have the ability to be limited in time, revoked or regenerated. It is a good security standard and removing it represents weakening. Clutter? LOL. A regular user wouldn't even be able to find it. Certainly, it does not pop up anywhere, one has to find it.
Click to expand...
Click to collapse
It is not something the average user should have access to, for several reasons. The TextSecure V2 protocol is NOT comparable with PGP/GPG because it has forward secrecy and deniability. The keys that are actually used to encrypt a message are not static as with PGP.
They are derived from the original keys and are changed with every message. No need to change them after X days/months/years.
Even if one key is intercepted, you would only be able to decrypt one message and not every message as it is the case with PGP.
If you get a new key, all your contacts get alerts that your key changed and that somebody may be listening in. That's not something the average user should be exposed to. If you think for whatever reason that you really want to do this, back up your conversations, uninstall TextSecure, install it again, import the backup and you have your new key.
Number three: Sideload or compiling: a regular user will do neither, he/she will simply download the app from the market, which means he has to have Google blobs. Or you are suggesting that users should download the app from the market and then remove GSF and other Googleapps? LOL again.
As I said earlier, Moxie's argument that allowing third party apps on your device is a greater security risk than having closed source blobs is wrong and grand BS (especially coming from someone who is considered a security expert). It is security through obscurity, which is no security at all. The value of his open source project is completely defeated by having closed source blobs by a known private branch of known three letter agencies.
Click to expand...
Click to collapse
Every average user has the google blobs, because they are preinstalled on nearly every phone and it's nearly unusable without them. This app is supposed to make encryption available to the masses.
Google may be undermined by your beloved three letter agencies, but it's not one of them. This is not to hide from them.
You have your threat model wrong.
No app alone can ever protect you from those agencies. They have hundreds of 0days for every platform and will simply own your Android, open source or not.
And this is not what TextSecure tries to do. They protect the content of every conversation with extremely strong encryption, no matter what the transport is. This does protect you from dragnet surveillance. But they can not protect you from someone who targets you and is willing to spend hundreds of thousands or millions to break into your operating systems.
If the NSA really wants you they get you, period. But TextSecure protects you from theives, cyber criminals and nearly everybody else who wants to read your messages.
You say you think the encrypted SMS mode was safe? With this your provider (and thus your government and every agency that wants it) has all the metadata. Who sent something to whom etc.
Google on the other hand has actually LESS meta data, because your phone sends the message to the TextSecure server, which relays the message to GCM. GCM then delivers the message. Because everything is encrypted none of the servers get contact data. But google only gets the receiver, not the sender. Your provider gets everything.
A global passive adversary may still do time corellation attacks, by listening who sends something when and who receives something at this time. After some sessions it's pretty clear who is talking to whom. It doesn't matter if Google is evil or not in this case. They get the metadata if they want to.
If you want protection against something like this take a look at pond, or meet i person: https://github.com/agl/pond
Now, these are facts. Let's get to opinions. I think that this deliberate weakening of security (again coming from a security expert) is a strong indication that development and/or developer has been compromised. And that is why I recommend to stay away from this app. But that is just my opinion, which is nonetheless based on facts.
Click to expand...
Click to collapse
As I explained there is no weakening whatsoever. Even if you consider google the adversary, they get less meta data than your SMS provider.
You can use this exactly as before without the google blobs if you want to.
They are actively working on a way to get away from the play store and GCM by building their own distribution method (which is finished, but not yet released, see #127 in their github) and implementing Websockets (server works, client is on the way).
Before you start slamming something you should really understand how it works, or ask if you understood it correctly.
lindworm said:
Do you even read what I write?
As I explained he does now work there any more.
You seem to have noticed that too:
Are you kidding me? How the flying **** did you get to this conclusion? The company that was bought by twitter was Whisper Systems.
They are publishing the new source under Open Whisper Systems. (none of those was ever called Whisper)
See the difference? They also state this here: http://support.whispersystems.org/customer/portal/articles/1474591-is-textsecure-owned-by-twitter-
So you are saying CyanogenMod is part of this grand conspiracy of yours? Come on...
It's a binary blob and it sends data to google, but you have no proof whatsoever if it records keystrokes. You can know if you want to tough. Decompile it and analyze it. I don't like binary blobs, but you can't just say they do something without having any proof. I may not be able to guarantee that they don't do something, because I have not personally decompiled and analyzed every bit of it, but until you have and have proof that it does do something you can't just claim it does.
It is not something the average user should have access to, for several reasons. The TextSecure V2 protocol is NOT comparable with PGP/GPG because it has forward secrecy and deniability. The keys that are actually used to encrypt a message are not static as with PGP.
They are derived from the original keys and are changed with every message. No need to change them after X days/months/years.
Even if one key is intercepted, you would only be able to decrypt one message and not every message as it is the case with PGP.
If you get a new key, all your contacts get alerts that your key changed and that somebody may be listening in. That's not something the average user should be exposed to. If you think for whatever reason that you really want to do this, back up your conversations, uninstall TextSecure, install it again, import the backup and you have your new key.
Every average user has the google blobs, because they are preinstalled on nearly every phone and it's nearly unusable without them. This app is supposed to make encryption available to the masses.
Google may be undermined by your beloved three letter agencies, but it's not one of them. This is not to hide from them.
You have your threat model wrong.
No app alone can ever protect you from those agencies. They have hundreds of 0days for every platform and will simply own your Android, open source or not.
And this is not what TextSecure tries to do. They protect the content of every conversation with extremely strong encryption, no matter what the transport is. This does protect you from dragnet surveillance. But they can not protect you from someone who targets you and is willing to spend hundreds of thousands or millions to break into your operating systems.
If the NSA really wants you they get you, period. But TextSecure protects you from theives, cyber criminals and nearly everybody else who wants to read your messages.
You say you think the encrypted SMS mode was safe? With this your provider (and thus your government and every agency that wants it) has all the metadata. Who sent something to whom etc.
Google on the other hand has actually LESS meta data, because your phone sends the message to the TextSecure server, which relays the message to GCM. GCM then delivers the message. Because everything is encrypted none of the servers get contact data. But google only gets the receiver, not the sender. Your provider gets everything.
A global passive adversary may still do time corellation attacks, by listening who sends something when and who receives something at this time. After some sessions it's pretty clear who is talking to whom. It doesn't matter if Google is evil or not in this case. They get the metadata if they want to.
If you want protection against something like this take a look at pond, or meet i person: https://github.com/agl/pond
As I explained there is no weakening whatsoever. Even if you consider google the adversary, they get less meta data than your SMS provider.
You can use this exactly as before without the google blobs if you want to.
They are actively working on a way to get away from the play store and GCM by building their own distribution method (which is finished, but not yet released, see #127 in their github) and implementing Websockets (server works, client is on the way).
Before you start slamming something you should really understand how it works, or ask if you understood it correctly.
Click to expand...
Click to collapse
"Decompile GSF"
You are kidding. Aren't you? If one can examine closed source the same way as open one, then all problems would be solved. And by the way, there would be no point in having proprietary software. Would it? Of course Java is easier to reverse engineer, but want to try Oracle's java?
"Google" Google has root access to your device: It can pull/install any application without you noticing it. They can install another version of TextSecure with backdoors. They can do whatever they want or told to. So, if you have Google, there is no point in any security at all. And when a developer forces users to have Google for his app to work, that's no security at all.
Cyanogenmode/Conspiracy? There is no conspiracy. The US has a law that requires providers to have back doors in their software/hardware for law enforcement, and there are wild claims (by those who know (and don't) what they are talking about) of TextSecure as "weapon" against this kind of surveillance. And that is pure bull. All that the app can provide is the false sense of security, while in reality making users more transparent to surveillance.
Phone service providers vs. internet: when you use Textsecure as a pure sms app, your provider gets gibberish, but they have no way of knowing what you are using. With GCM/GSF/Googleplay, they know exactly what you are doing, as you are marked as using this particular app. So, Moxie is making life of "survaillors" much easier.
Thanks for telling me to uninstall the app if I want to generate new key. So, if I do it this way, you think my contacts won't receive a message that my key has changed?
Here is how I began to suspect foul play: First I noticed the app wanted access to the internet, then I discovered that I can no longer generate a new key, then I went to read about F-droid/Whisper problems. Then I read that he wants the app be available through Google only, because he cares about security and does not want users to allow third party apps (BS). Then I read about feds harassment. You think the 3 letter agencies wouldn't like to have him?
In my view, Moxie's arguments no longer make sense. And by the way, when he is against the wall, he tells you to create a world wide push service - alternative to GCM. LOL.
For me that's enough to stay away from the app. Others will decide accordingly...
Does anybody work on an alternativ push service in order to replace hard requirement on Google services for TextSecure, Redphone and lots of other useful apps?
I understand that GAPPS are needed to run textsecure.
Is it possible/ has anyone succeed to get it to run with the no GAPPS apps such as the blank store etc or is the app relying too much on google infrastructure?
i can use textsecure sms without internet. besides registering with push is not mandatory at all so the crash you've experienced must be a bug in the version of textsecure you're using. also why compare it to pgp/gpg? textsecure uses otr with improvements to deniability and forward secrecy. also textsecure supports mms (which uses internet).
if you're really that paranoid, avoid android at all and stop spreading FUD claiming it to be fact. i don't find the statement factual at all. it lacks any evidence (show us the code with the backdoor first).
and also avoid openguardian project too as they conspire with textsecure since they are recommending it.
and by the way, whisper and openwhisper are different.
It really is ashamed when misinformed people comment on things they do not have enough information to intelligently speak about. Especially when it discourages people from using an application that is one of the only current means of communicating over SMS in a secure manner. Is it perfect? Certainly not... Security and encryption are never perfect, and there will always be flaws to be found, but to insist that someone such as Moxie Marlinspike is somehow working against the security researcher community in some undercover role as an agent of the government or some corrupt company is really insulting. If you have some absolute proof, or even a reasonable solid suspicion, please share it, but otherwise do not taint these incredible people with false accusations. Learn a bit about encryption, reverse engineering, and packet inspection, and then come back and give an intelligent analysis of your findings of the application you suspect to be playing some nefarious role. Until then, your accusations are completely unfounded and damaging to the community as a whole. There are many people who have worked hard to make this product a reality, and I believe they should be praised for their efforts. Obviously these are my own opinions, and you are free to dismiss them outright as you have done to others in previous posts. In addition, I realize I am not an active member of the xda community, but I am an active member of the security/reverse engineering community. My job and nearly all of my free time is spent reverse engineering software and I see no basis for your accusations.
Here is more update on Textsecure: there was a major vulnerability found last October-November. And Moxie's response (not surprisingly) - fixing "feels pretty cumbersome" and "I dunno."
Also, Open Whisper is now accepted into the family of such a bastion of privacy, as Facebook (kids love it, NSA approves). So, If you had any doubt about this app before, now you can sleep well at night (sarcasm).
https://moderncrypto.org/mail-archive/messaging/2014/001029.html
https://moderncrypto.org/mail-archive/messaging/2014/001030.html
To those who like to attack the messenger ( I call them Google thugs or pacifier babies). One says decompile GSF, the other - false accusations and absolute proof?! Wake up and get the pacifier out of your mouth. There is no such thing in real life. I give you the dots, you can't connect them with the pacifier in your mouth.
Here is some more damning evidence that Textsecure is a totally compromised project no longer to be trusted: during 2013-2014 Open Whisper Systems received over $1.3 mln from BBG, which is an arm of US Government and its 3-letter-agencies.
http://pando.com/2015/03/01/internet-privacy-funded-by-spooks-a-brief-history-of-the-bbg/
So, Moxie, it appears, has turned from someone who was harrased by TSA in airports (presumably for a failure to cooperate with the government) to a receipient of major funds from the same government. I am not even talking about him getting a once in a life-time project to work on "securing" Facebook's What's up application. Pitty and shame...
Replacement for Textsecure
Here is a pure sms app, which replaces compromised Textsecure, as well as stock messaging. There is no over the internet messaging, no google binaries and no Google Services Framewor all closed sourse. In addition, starting from version 2.7, textsecure no longer encrypts SMS. Pitty.
Here is the latest version: http://forum.xda-developers.com/android/apps-games/sms-secure-aes-256-t3065165
Short version: I programmed a Windows 8 Oauth app. I didn't know where to post this, but it's mostly done in javascript and HTML so I figured this forum might be best. If others have time, I'd really appreciate it if someone would audit my code. Due to the nature of the amount my request, I thought it would be best to post a link to the GitHub repo. If this is wrong, please correct me.
GitHub: https://github.com/mepis/Windows8OauthAuthenticator
Long Version: I use 2-step for a lot of my accounts. The problem is, I'm lazy. I don't feel like getting up to get my phone after I set it down at night. I wanted a metro Oauth app for Windows 8. I looked on the store, but didn't recognize any of the developers. Due to the nature of Oauth, I choose to err on the side of caution and not use the apps. I'm not saying that other devs aren't well intentioned and good devs. I'm just saying that it's a better idea in the name of security that I not use the apps if I can't verify anything. So I decided to write my own.
That leaves another issue though. Due to the nature of Oauth, the token device shouldn't be on the same device you're putting passwords in. I'm choosing to ignore this a bit. I do recognize that tokens shouldn't be stored in plain text though in the Windows storage space. Instead, I push and pull the token from the Windows Credential Manager and the password vault.
I was thinking of running the tokens, labels, and account names through an AES algorithm and then storing that information in the credential manager. This would require a user password on opening the app though. I'm not sure I want to go that route yet, though it would be easily implemented later on.
The mission of this app is simple. I want to offer an Oauth app that is open source and able to be audited by the general public. I want others to have access to a free tool that they can trust and review. I will never charge for this app nor ask for donations. It's also posted under the GNU version 3 license.
At some point, I am thinking about porting this app to Windows Phone.
I'm very much a amateur developer though. I was hoping that others could audit my app, offer suggestions, and point out mistakes. I very much appreciate any help or time that any person is willing to offer.
While you may well get some takers, and some of them might even know what they're doing, you realize you're asking for something that is usually done by people who do this stuff professionally for hundreds of dollars per hour, right? It's like writing up a legal contract and posting it online and saying "do you think this will hold up in court?"
OK, training to be a security engineer doesn't take as long as training to be a lawyer. But there's *more* lawyers than there are security engineers, and our time is very much in demand (yes, I'm a security engineer; no, I will not audit your code for free unless I expect to have a use for it personally).
I'm not even sure what you mean by "OAuth app". OAuth is a standardized protocol (v2.0, RFC 6749, is more accurately described as a framework) for delegated authentication. For example, you've seen how a lot of web sites let you sign in using your Facebook account? That's because they use Facebook as an OAuth provider. The website delegates the responsibility of authenticating users to Facebook, which is handy for them because they don't have to handle passwords and so forth, handy for the user because many users already have FB accounts, and handy for FB because they gain information about what kinds of sites you visit and can use that to target ads. It also has downsides, of course; the OAuth client (web site) has to trust that FB knows what they're doing and to remain available, the user gives FB info they might not want FB to have and also ends up essentially re-using passwords across sites (a bad idea), and FB bears the cost and responsibility of managing all those logins.
Now, to make any authentication scheme (including but not limited to OAuth) stronger, you can multi-factor authentication (sometimes called two-factor auth or 2FA). The most common way of doing that is using Time-based One Time Password (TOTP, standardized as RFC 6238) security tokens, either in small hardware devices or in mobile apps. Is that what this is supposed to be? Because... that has nothing to do with OAuth.
I have a hard time imagining a situation in which I'd use a TOTP generator written by somebody who didn't know the difference between TOTP and OAuth.
Well, your response thus far has been excellent (I'm not being sarcastic). I need to read more about Oauth then. I must have my definitions and understanding a bit confused.
In actuality, to phrase it better, the application would be a TOTP app then - like Google Authenticator. I used Javascript provided by Google for the TOTP generation. The app itself is rather simple. My biggest concern though is the safety of the tokens. I used Windows Credential Manager to store the tokens on the device. I couldn't find much information about the security of Windows Credential Manager though. That's my biggest concern.
Other than that, thanks for the information. I'm going to do some more reading.
For what it's worth (and without having read your code), it sounds like you're doing OK; TOTP generators are not complex by themselves, and usually the only threat to them is in the secret storage (which you're addressing). Of course, most of them offer things like QR code scanning (as a way to load secrets more easily) and I don't know if you have anything like that or whether there are any security pitfalls there.