Database Users

Android Security: A neglected subject (long)

First of all: I'm an OSS advocate and love the idea of open source. Don't forget that while reading this.
Some 2 month ago, I got myself a Galaxy S. It's not exactly cheap, but on the other side, it's really good hardware. This thread is not about Samsung or the Galaxy S. It's about the missing parts of android security.
We all know it from our home computers: Software sometimes has bugs. Some just annoy us, others are potentially dangerous for our beloved data. Our data sometimes gets stolen or deleted due to viruses. Viruses enter our machines by exploiting bugs that allow for code execution or priviledge escalation. To stay patched, we regularly execute our "apt-get update;apt-get dist-upgrade" or use windows update. We do this to close security holes on our systems.
In the PC world, the software and OS manufacturers release security bulletins to inform users of potentially dangerous issues. They say how to work around them or provide a patch.
How do we stay informed about issues and keep our Android devices updated?
Here's what Google says:
We will publicly announce security bugs when the fixes are available via postings to the android-security-announce group on Google Groups.
Click to expand...
Click to collapse
Source: http://developer.android.com/guide/appendix/faq/security.html#informed
OK, that particular group is empty (except for a welcome post). Maybe there are no bugs in Android. Go check yourself and google a bit - they do exist.
"So why doesn't Google tell us?", you ask. I don't know. What I know is that the various components of Android (WebKit, kernel, ...) do have bugs. There's nothing wrong with that BTW, software is made by people - and people make mistakes and write buggy code all the time. Just read the changelogs or release notes.
"Wait", I head you say, "there are no changelogs or release notes for Android releases".
Oh - so let's sum up what we need to stay informed about security issues, bugs and workarounds:
* Security bulletins and
* Patches or Workaround information
What of these do we have? Right, nada, zilch, rien.
I'll leave it up to you to decide if that's good common practise.
"But why is this important anyway", you ask.
Well, remember my example above. You visit a website and suddenly find all your stored passwords floating around on the internet. Don't tell me that's not possible, there was a WebKit bug in 2.2 that did just that. Another scenario would be a drive-by download that breaks out of the sandbox and makes expensive phone calls. Or orders subscriptions for monthly new ringtones, raising your bill by orders of magnitute. Or shares your music on illegal download portals (shh, don't tell the RIAA that this is remotely possible).
The bug is probably fixed in 2.2.1 - but without changelogs we can't be sure.
But that's not all - there's a second problem. Not only are we unaware of security issues, we also don't have automated update mechanisms.
We only receive updates when our phone's manufacturers release new firmware. Sadly, not all manufacturers support their phones in the long run.
In the PC world, most Distros have a central package management - that Google forgot to implement in Android. Agreed, some phones can receive OTA updates, but that depends on the carrier. And because of the differences in Android versions it's not possible to have a central patch management either. So we do not know if our Android devices might have security issues. We also have no easy way to patch them.
Perhaps you knew this before, then I apologize for taking your time.
What do YOU - the computer literate and security aware XDA users - think about this? Do you think that's a problem? Or would you rather say that these are minor problems?

Very intresting, thanks! The update problem should be fixed with the next release, no more custom UIs and mods from phone manufacturers,at least google said that
Sent from my Nexus One using XDA App

Excellent post and quite agree with you. The other significant problem looming is the granularity (or rather, lack thereof) in app permissions which can cause problems you describe without bugs and exploits. I install an app that does something interesting with contacts and also has internet access to display ads. How do I know that my contacts are not encrypted, so making sniffing useless, and beamed back to mummy? Nothing other than blind trust!
I love Android but it's an accident waiting to happen unless the kind of changes you advocate are implemented and granularity of permissions significantly increased. I don't like much about Apple but their walled garden app store is something they did get right although IMHO, they also abuse that power to stifle competition. Bring out the feds!

simonta said:
The other significant problem looming is the granularity (or rather, lack thereof) in app permissions [...]
How do I know that my contacts are not encrypted, so making sniffing useless, and beamed back to mummy? Nothing other than blind trust!
Click to expand...
Click to collapse
I agree, although I'm not sure that less experienced users might have difficulties with such options.
simonta said:
I love Android but it's an accident waiting to happen
Click to expand...
Click to collapse
Sad but true. I'm just curious what Google will do when the first problems arise and the first users will have groundshaking bills.
If that happens to just a few users, it'll get a kind media coverage Google surely won't like.

I've seen quite a few android exploits posted on bugtraq over the years. It's a high-volume email list, but with some filtering of stuff you don't care about, it becomes manageable. It's been around forever and is a good resource if you want the latest security news on just about anything computer related.
http://www.securityfocus.com/archive/1/description

People are bashing a lot about the Android security model but the truth is you can never have 100% protection with ANY solution.
Apple is not allowing any app in their store. Fine. but mostly they are only filtering out apps that crash, violate some rules or they just don't like them or whatever. but they can never tell what an app is really doing. Therefore they would neeed to reverse-engineer every app they get etc. That's just impossible considering the amount of apps....
Speaking again of Android. I think the permission model is not bad. I mean, no other OS got such detailed description about what an app can do or not. But unfortunately it can only filter out very conspicuous apps, i.e. a Reversi game asking for your location and internet access. But then you never know... if the app is using ads it requires location and internet access, right? so what can you do?

RAMMANN said:
Apple is not allowing any app in their store. Fine. but mostly they are only filtering out apps that crash, violate some rules or they just don't like them or whatever. but they can never tell what an app is really doing. Therefore they would neeed to reverse-engineer every app they get etc. That's just impossible considering the amount of apps....
Click to expand...
Click to collapse
Not really, they do blackbox testing and let the apps run on emulated devices they then check if the app "behaves" as desired...

Of course you can't get 100% security and I don't think that's what we're saying, but there is a lot you can do.
Take for example internet access which is the biggest worry I have. The only reason most apps request internet access is to support ads. I now have a choice to make, don't use the app or trust it. That simple, no other choice.
If I installed an app that serves ads but did not have internet access, then the only way that app can get information off my phone is to use exploits and I'm a lot more comfortable knowing that some miscreant needs to understand that than the current situation where some script kiddy can hoover up my contacts.
However, if internet access and ad serving were separate permissions, you could in one hit address, taking a wild guess, 90% of the risk from the wild west that is Marketplace. With a bit more design and work, it would be possible to get the risk down to manageable and acceptable levels (at least for me).
I absolutely agree with you on Apple, one of the main reasons that I chose a Desire instead of an iPhone, but the Android approach is too far the other way IMHO.
Just my tuppence, in a hopeless cause of imagining someone at Google paying attention and thinking you know what, it is an accident waiting to happen.

marty1976 said:
Not really, they do blackbox testing and let the apps run on emulated devices they then check if the app "behaves" as desired...
Click to expand...
Click to collapse
Well, so why did a tethering app once make it into the appstore?
Also I think there are many possibilities for an app to behave normal, and just start some bad activity after some time. Wait a couple months until the app is spread around and then bang. Or remotely launch some action initiated through push notifications etc.
If there is interest, then there is always a way....

simonta said:
However, if internet access and ad serving were separate permissions, you could in one hit address, taking a wild guess, 90% of the risk from the wild west that is Marketplace. With a bit more design and work, it would be possible to get the risk down to manageable and acceptable levels (at least for me).
Click to expand...
Click to collapse
I agree that a seperate permission for ads would be a good thing.
But there are still many apps which need your location, contacts, internet access.... all the social media things nowadays. And this is where the whole thing will be going to so I think in the future it will be even harder to differenciate.

Getting back on topic: I just read that Windows 7 Phone will get updates and patches like desktop windows. That means patchday once a month plus when urgency is high...

simonta said:
However, if internet access and ad serving were separate permissions, you could in one hit address, taking a wild guess, 90% of the risk from the wild west that is Marketplace. With a bit more design and work, it would be possible to get the risk down to manageable and acceptable levels (at least for me).
Click to expand...
Click to collapse
But, how do you distinguish them? Today, (as a developer) I can use any ad-provider I want. In order to distinguish ads from general internet access, the OS would need one of:
A Google-defined ad interface, which stifles "creativity" in ad design. Developers would simply ignore it and do what they do now as soon as their preferred ad-provider didn't want to support the "official" ad system or provided some improvement by doing so.
An OS update to support every new ad-provider (yuck^2).
Every ad-provider would have to go through a Google whitelist that was looked up on the fly (increased traffic, and all ads are now "visible" to Google whether Google is involved in the transaction or not). This would also make ad-blocking apps harder to implement since Google's whitelisting API might not behave if the whitelist was unavailable. On the upside, it would make ad-blocking in custom ROMs be trivial.
Even if Google did one of these things, it still wouldn't provide any real increase in privacy or security. The "ad service" would still need to deliver a payload from the app to the service (in order to select ads) and another from the service to the app (the ad content). Such a mechanism could be trivially exploited to do anything that simple HTTP access could provide.

http://code.google.com/p/android/issues/list
issues submitted are reviewed by google employed techs... they tell you if you messed up and caused the issue or if the issue will be fixed in a future release or whatever info they find.
probably not the best way to handle it but its better then nothing.

twztdwyz said:
http://code.google.com/p/android/issues/list
Click to expand...
Click to collapse
Knew that bug tracker, but the free tagging aka labels isn't the best idea IMHO.
You can't search for a specific release, for example...
twztdwyz said:
probably not the best way to handle it but its better then nothing.
Click to expand...
Click to collapse
Ack, but I think Google can do _much_ better...

Two more things to have in mind:
1. I doubt that many Android users bother much about what permissions they give to an app.
2. Using Google to sync your contacts and calendar (and who knows what else), is a bad, bad idea.

Android 2.2.2 Security

2.2.2 has a security fix
http://www.engadget.com/2011/03/02/google-spikes-21-malicious-apps-from-the-market-with-big-downloa/
thoughts?

My thoughts are simple: Sprint needs to get its **** together and release an official 2.3 release. And Google needs to consider some sort of authentication program for apps to be distributed in the Market.
Certainly don't want to cut the independent developer community off, but it shouldn't be their responsibility to release new versions of essential operating software that contain fixes that disable malicious exploits. They are here to enhance our user experience.
The manufacturers need to be concerned about what the deleterious effects of outdated software can open their networks to. After all, these apps had full internet access, as I've heard. Who knows if, say a DDOS attack (or something worse), could be possible using phones, and what kind of effects that could have on the stability of the entire Sprint network.
As for Google, I'm not suggesting that the Market be completely walled-off, but maybe having something like "Google Approved" or "Verified Secure" or something, would give us users more confidence that apps come from verified and vetted sources. We could still install things not verified -- at our own risks -- but at least we'd have a choice and be able to proceed with better, more complete information.

TonyArmstrong said:
My thoughts are simple: Sprint needs to get its **** together and release an official 2.3 release. And Google needs to consider some sort of authentication program for apps to be distributed in the Market.
Certainly don't want to cut the independent developer community off, but it shouldn't be their responsibility to release new versions of essential operating software that contain fixes that disable malicious exploits. They are here to enhance our user experience.
The manufacturers need to be concerned about what the deleterious effects of outdated software can open their networks to. After all, these apps had full internet access, as I've heard. Who knows if, say a DDOS attack (or something worse), could be possible using phones, and what kind of effects that could have on the stability of the entire Sprint network.
As for Google, I'm not suggesting that the Market be completely walled-off, but maybe having something like "Google Approved" or "Verified Secure" or something, would give us users more confidence that apps come from verified and vetted sources. We could still install things not verified -- at our own risks -- but at least we'd have a choice and be able to proceed with better, more complete information.
Click to expand...
Click to collapse
+1 but i also think they should make an official malware scanner.

Rydah805 said:
+1 but i also think they should make an official malware scanner.
Click to expand...
Click to collapse
This.^^^^
I'm an Android convert (from iPhone), and my great fear is that the very openness we enjoy could expose us to very nasty ****. I don't wanna be locked down, but I do want some manner of enhanced security.
That malware scanner in combo with some sort of developer authentication and/or verification program would be excellent.

[APP] FREE Anti virus

http://forum.xda-developers.com/showthread.php?t=1453695

Why are you creating 2 topics about it?
Had you tested it? How it compare to theoretically best Zoner Antywirus? Tell us some more, than posting links - this is kind of flooding.
For me, this program won't beat Zoner.. for now.
Anyway, I'll test it

Rayman96 said:
Why are you creating 2 topics about it?
Had you tested it? How it compare to theoretically best Zoner Antywirus? Tell us some more, than posting links - this is kind of flooding.
For me, this program won't beat Zoner.. for now.
Anyway, I'll test it
Click to expand...
Click to collapse
sorry if i did hurt you. well i was a beta tester for the app. it did performed well for me, besides comodo is a reputed company after all and they are standing for free softwares.
I posted the links cause it contains all the details of the software, details about the company etc, i thought its better than i explain those details.
about double posting, the one i posted is in the general section is for all to see. The second is for my fellow lgp500 users, where i really belogs. i hope i am clear enough. no harm ment

Best free antivirus is your brain - never install app without good amount of comments about app.

AdvDretch said:
Best free antivirus is your brain - never install app without good amount of comments about app.
Click to expand...
Click to collapse
Who in this world has time to read all that? Have you ever tried to read Google’s conditions and policies while creating a Google account? Certainly the answer would be ‘NO’. Do you know that Google had 60 different policies that helped them to collect data from your personal Gmail and other Google apps? Now do you know that they had merged all these in to one policy?
Google will know more about you than your wife does. Everything across your screens will be integrated and tracked. Google noted that it collects information you provide, data from your usage, device information and location. Unique applications are also noted. Sure you can use Google’s dashboard and ad manager to cut things out, but this policy feels Big Brother-ish. Google is watching you as long as you are logged in. It’s also unclear whether this privacy policy move will be considered bundling in some way by regulators. This unified experience hook appears to be at least partially aimed at juicing Google+. Google responded with clarification: Google noted that it already has all that data, but it’s now integrating that information across products. It’s a change in how Google will use the data not what it collects. In other words, Google already knows more about you than your wife.( not my comment go read this.... http://m.zdnet.com/blog/btl/googles-new-privacy-policy-the-good-bad-scary/67893)
Now my question is whether Google is good or bad? Do you need Droidwall to defend your privacy? Or do you still believe in your Brain(better do not believe in brain but use it to think rationally)?
Conclusion: we need a new definition to “virus”...My contribution is Anything that steals your private data is a virus.( no flames needed, no harm meant...just my thought about the relevancy of protective apps like Droidwall, comodo, avg, etc. ...etc)

,do we realy need anti virus?,

algie17 said:
,do we realy need anti virus?,
Click to expand...
Click to collapse
You dont need one
Sent from my LG-P500 using XDA Premium App

josinpoul's mean run anti virus before creating Google account
And if too don't have anti virus then don't use Google. Josin your explanation is wrong. Brain and antivirus both useful.

No need for 2 topics about one thing but thanks for sharing!!!

http://ca.reuters.com/article/technologyNews/idCATRE81N1T120120224
By Jim Finkle
BOSTON (Reuters) - Cybersecurity experts have uncovered a flaw in a component of the operating system of Google Inc's widely used Android smartphone that they say hackers can exploit to gain control of the devices.
Researchers at startup cybersecurity firm CrowdStrike said they have figured out how to use that bug to launch attacks and take control of some Android devices.
CrowdStrike, which will demonstrate its findings next week at a major computer security conference in San Francisco, said an attacker sends an email or text message that appears to be from a trusted source, like the user's phone carrier. The message urges the recipient to click on a link, which if done infects the device.
At that point, the hacker gains complete control of the phone, enabling him or her to eavesdrop on phone calls and monitor the location of the device, said Dmitri Alperovitch, chief technology officer and co-founder of CrowdStrike.
Google spokesman Jay Nancarrow declined comment on Crowdstrike's claim.
Alperovitch said the firm conducted the research to highlight how mobile devices are increasingly vulnerable to a type of attack widely carried out against PCs. In such instances, hackers find previously unknown vulnerabilities in software, then exploit those flaws with malicious software that is delivered via tainted links or attached documents.
He said smartphone users need to prepare for this type of attack, which typically cannot be identified or thwarted by mobile device security software.
"With modifications and perhaps use of different exploits, this attack will work on every smartphone device and represents the biggest security threat on those devices," said Alperovitch, who was vice president of threat research at McAfee Inc before he co-founded CrowdStrike. Researchers at CrowdStrike were not the first to identify such a threat, though such warnings are less common than reports of malicious applications that make their way to online websites, such as Apple's App Store or the Android Market.
In July 2009, researchers Charlie Miller and Collin Mulliner figured out a way to attack Apple's iPhone by sending malicious code embedded in text messages that was invisible to the phone's user. Apple repaired the bug in the software a few weeks after the pair warned it of the problem.
The method devised by CrowdStrike currently works on devices running Android 2.2, also known as Froyo. That version is installed on about 28 percent of all Android devices, according to a Google survey conducted over two weeks ending February 1.
Alperovitch said he expects to have a second version of the software finished by next week that can attack phones running Android 2.3. That version, widely known as Gingerbread, is installed on another 59 percent of all Android devices, according to Google.
CrowdStrike's method of attack makes use of a previously unpublicized security flaw in a piece of software known as webkit, which is built into the Android operating system's Web browser.
Webkit is also incorporated into other software programs, including Google's Chrome browser and the Apple iOS operating system for the iPhone and iPad.
CrowdStrike said it had not attempted to create software to attack iOS devices or the Chrome browser.

Ok, now a group of hackers control 500000000 devices... an antivirus will slow the phone down more than a hacker trying to run a phone from another continent over your 2G network... just think about it... how can your screen be monitored over 3G in real-time? It can't be done on my 5Mbps PC...
And if you turn data off, then 1GB of data will be sent to google when you turn it on??? Think logic...(where the f**k do you store that??? I think the effect will be noticed right away, and the attacker has no time to take control, unless you are stupid enough to see a 1GB file and not suspect anything...) PCs have real-time protection, but that is because there are terrible threats out there, and they are optimized, they don't slow down... on your phone, you will regret having a phone for 2 years running like **** and then dropping in water, while you could have best performance in those 2 years...
We are not windows, but we are android, and it is the most unsafe mobile OS, if you want a safe one, get from apple... just 2x price at ½ quality...
Sent from my LG-P500

well i use avast antivirus
but not for scanning viruses
but rather for anti-theft feature and firewall(blocking apps)
and isnt android a java based OS ??
im sure there are not many virus's
that can cause heavy damage

Vulnerability Allows Attackers to Modify Android Apps Without Breaking Their Signatur

Vulnerability Allows Attackers to Modify Android Apps Without Breaking Their Signatures
This might be the reason why the new MF2 and ME6 are not downgradable and why the 4.2.2 update was delayed.
Source->http://www.cio.com/article/735878/V...ndroid_Apps_Without_Breaking_Their_Signatures
IDG News Service — A vulnerability that has existed in Android for the past four years can allow hackers to modify any legitimate and digitally signed application in order to transform it into a Trojan program that can be used to steal data or take control of the OS.
Researchers from San Francisco mobile security startup firm Bluebox Security found the flaw and plan to present it in greater detail at the Black Hat USA security conference in Las Vegas later this month.
The vulnerability stems from discrepancies in how Android apps are cryptographically verified, allowing an attacker to modify application packages (APKs) without breaking their cryptographic signatures.
When an application is installed and a sandbox is created for it, Android records the application's digital signature, said Bluebox Chief Technology Officer Jeff Forristal. All subsequent updates for that application need to match its signature in order to verify that they came from the same author, he said.
This is important for the Android security model because it ensures that sensitive data stored by one application in its sandbox can only be accessed by new versions of that application that are signed with the original author's key.
The vulnerability identified by the Bluebox researchers effectively allows attackers to add malicious code to already signed APKs without breaking their signatures.
The vulnerability has existed since at least Android 1.6, code named Donut, which means that it potentially affects any Android device released during the last four years, the Bluebox researchers said Wednesday in a blog post.
"Depending on the type of application, a hacker can exploit the vulnerability for anything from data theft to creation of a mobile botnet," they said.
The vulnerability can also be exploited to gain full system access if the attacker modifies and distributes an app originally developed by the device manufacturer that's signed with the platform key -- the key that manufacturers use to sign the device firmware.
"You can update system components if the update has the same signature as the platform," Forristal said. The malicious code would then gain access to everything -- all applications, data, accounts, passwords and networks. It would basically control the whole device, he said.
Attackers can use a variety of methods to distribute such Trojan apps, including sending them via email, uploading them to a third-party app store, hosting them on any website, copying them to the targeted devices via USB and more.
Some of these methods, especially the one involving third-party app stores, are already being used to distribute Android malware.
Using Google Play to distribute apps that have been modified to exploit this flaw is not possible because Google updated the app store's application entry process in order to block apps that contain this problem, Forristal said. The information received by Bluebox from Google also suggests that no existing apps from the app store have this problem, he said.
However, if an attacker tricks a user to manually install a malicious update for an app originally installed through Google Play, the app will be replaced and the new version will no longer interact with the app store. That's the case for all applications or new versions of applications, malicious or non-malicious, that are not installed through Google Play, Forristal said.
Google was notified of the vulnerability in February and the company shared the information with their partners, including the members of the Open Handset Alliance, at the beginning of March, Forristal said. It is now up to those partners to decide what their update release plans will be, he said.
Forristal confirmed that one third party device, the Samsung Galaxy S4, already has the fix, which indicates that some device manufacturers have already started releasing patches. Google has not released patches for its Nexus devices yet, but the company is working on them, he said.
Google declined to comment on the matter and the Open Handset Alliance did not respond to a request for comment.
The availability of firmware updates for this issue will differ across device models, manufacturers and mobile carriers.
Whether a combination of device manufacturers and carriers, which play an important role in the distribution of updates, coincide to believe that there is justification for a firmware update is extremely variable and depends on their business needs, Forristal said. "Ideally it would be great if everyone, everywhere, would release an update for a security problem, but the practical reality is that it doesn't quite work that way, he said."
The slow distribution of patches in the Android ecosystem has long been criticized by both security researchers and Android users. Mobile security firm Duo Security estimated last September, based on statistics gathered through its X-Ray Android vulnerability assessment app, that more than half of Android devices are vulnerable to at least one of the known Android security flaws.
Judging by Android's patch distribution history so far, the vulnerability found by the Bluebox researchers will probably linger on many devices for a long time, especially since it likely affects a lot of models that have reached end-of-life and are no longer supported.
Click to expand...
Click to collapse

I really thought more people would be interested in knowing this. I would really like to know what you guys think about this.

Key phrase here is "for apps not installed through the google store". Hence not an issue for a large fraction of users. Total case of FUD. Someone must be wanting to sell some av software.
Sent from my GT-N7100 using Tapatalk 4 Beta

Kremata said:
I really thought more people would be interested in knowing this. I would really like to know what you guys think about this.
Click to expand...
Click to collapse
Well, X-Ray scanner either does not detect this latest security flaw or N7100 (as of DM6) is allready patched.

Kremata said:
I really thought more people would be interested in knowing this. I would really like to know what you guys think about this.
Click to expand...
Click to collapse
This is the first link I found for XDA on this.
I think it's not that interesting because it's old, old news and exactly why it's being touted as a "new" discovery is beyond me, it's far from new.
We here at XDA have been using this method for years to modify stock Android and OEM system apps with great success. Here's an example by me from 2011: http://forum.xda-developers.com/showthread.php?t=994544 there's a literally hundreds of examples all over XDA.
The real question here is how Bluebox security got everybody to act as a PR machine for them. If they turn up at Black Hat with this "amazing discovery" they're going to get laughed off the stage.

djmcnz said:
This is the first link I found for XDA on this.
I think it's not that interesting because it's old, old news and exactly why it's being touted as a "new" discovery is beyond me, it's far from new.
We here at XDA have been using this method for years to modify stock Android and OEM system apps with great success. Here's an example by me from 2011: http://forum.xda-developers.com/showthread.php?t=994544 there's a literry hundreds of examples all over XDA.
The real question here is how Bluebox security got everybody to act as a PR machine for them. If they turn up at Black Hat with this "amazing discovery" they're going to get laughed off the stage.
Click to expand...
Click to collapse
Ahh! Thats the answer I was waiting for (and from a Recognized Developer). I knew XDA Devs were using this method. My new question is.. If they fix it will it be harder to create Mods? Will it slow down development?

Shouldn't this be posted in the generals forum?

Kremata said:
If they fix it will it be harder to create Mods? Will it slow down development?
Click to expand...
Click to collapse
I suspect so. If they fix it properly it would become impossible to change any aspect of the app without signing it again. If you wanted to maintain compatibility with the original then you'd need the developer's keys.
At the moment really only the manifest and some metadata within the apk is signed, if they extended that to the entire contents of the apk many mods (think themes for stock Google apps etc) are screwed unless users are happy to relinquish Play Store links and updates (i.e. backward compatibility).
Google may not go this far and may only choose to authenticate the code (smali) rather than all of the apk contents (graphics, strings etc), this approach would leave room for some mods to survive. Remains to be seen.

99% android devices vulnerable. Desire Z too.

The vulnerability involves discrepancies in how Android applications are cryptographically verified & installed, allowing for APK code modification without breaking the cryptographic signature.
proof: http://bluebox.com/corporate-blog/bluebox-uncovers-android-master-key/
ROM: gen.y_vision_WWE_2.42.405.4_R1.3-EXT4 vulnerable.
I think, HTC will not update stock ROM, can i hope what xda developers will be update her ROMs?

As I understood this news, it must be user triggered. Someone must choose to install such an apk.
And most people that install from apk dont check the author of files anyways, so no big thing.
And if Playstore is in any way affected now or not, I do not know because I do not use it.
It always depends on how the stores/sites you get your apps from maintain a consistent trust relationship to app authors.
Would really be nice if a "Who is who and who made what" menu would be included into cm or aosp.

I think you are more or less safe when you stick to commonly used apps in the Play store. OTOH I had a app once that had notification ads in it, kind of spam.

The new ILWT CM7 supposedly has it fixed. Build 879

Unfortunately, Android accounts for 79% of phone malware and has become the new prime target. This really makes me think that having a good AV on my phone is important, yet how many of us don't? (including me!) Any recommendations that anyone has tested?

Come on, had anyone been hacked or smth?
We should just check software before installing, but i suppose google does it for us.