Feuerfest

Photo by Pixabay: https://www.pexels.com/photo/qr-code-on-screengrab-278430/

As I'm preparing to update the firmware on my mobile, updating the ROM and rooting it I'm currently in the "Backup, document and save everything" phase. This means I'm checking that I can backup and restore all my Two-Factor authentication codes (2FA) properly.

Partly because I didn't backup the enrolment QR-Codes for every service I signed up over the years. Documenting the otpauth:// URL and/or the initial QR-Code is still the best way. Then it doesn't even matter if all your devices get lost. You can just enter the information in your 2FA app use the code to sign in and then disable 2FA and re-enable it to invalidate the old secret. Locking out anyone else possibly using your devices/accounts.

As I played around a little bit with 2FA apps over the time I got three apps installed:

• Google Authenticator
• Aegis Authenticator
• FreeOTP

And here my problems are starting.

Google Authenticator: Only allows you to export your entries in a QR-Code. (Beside the Cloud sync, but whoever uses this hasn't properly understood 2FA in my personal opinion..)

Aegis Authenticator: Allows the export in un-/encrypted clear text. Even with proper otpauth:// URLs. Nice!

FreeOTP: Offers exporting your entries in an externalBackup.xml called file which contains JSON structured data!? Okay.. The secrets are encrypted with the password you chose when you installed the app. It cannot be changed or retrieved otherwise afterwards so I hope you remember it. 😉

There is a discussion on GitHub about how to decrypt that file, extract the secrets and build proper otpauth:// URLs, but that solution didn't work for me.

I only got the following error message:

user@host:~$ python3 freeotp.py
Traceback (most recent call last):
  File "/home/user/freeotp.py", line 26, in <module>
    tree = ET.parse("externalBackup.xml")
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/lib/python3.11/xml/etree/ElementTree.py", line 1218, in parse
    tree.parse(source, parser)
  File "/usr/lib/python3.11/xml/etree/ElementTree.py", line 580, in parse
    self._root = parser._parse_whole(source)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
xml.etree.ElementTree.ParseError: not well-formed (invalid token): line 1, column 0

Anyway, after deleting myself from several services in the last years only two entries were still needed and it was easier and faster to just disable 2FA on these accounts. So that is what I did.

EDIT November 2025: I only learned recently that Aegis supports the import of entries from FreeOTP and FreeOTP+ (along many other 2FA apps for that matter..). Either by importing the entries directly from the FreeOTP apps (requires root rights) or by exporting the entries from FreeOTP and importing the generated externalBackup.xml into Aegis. Technically this isn't a pro-argument for FreeOTP but rather another for Aegis. But it can also help FreeOTP users so...

Use the 3 vertical dots menu in the top-right of Aegis, choose Settings and navigate to: Imports and exports. There you have two options to import from file or from an app. EDIT END.

Using Aegis Authenticator to migrate to any 2FA app of choice

Regarding my entries in the Google Authenticator I generated the QR-Code and scanned that with Aegis Authenticator. Aegis properly imported all entries and the generated 2FA tokens were correct when I checked them against Google Authenticator.

As Aegis allows me to export everything in clear text I can use that to migrate to any 2FA app of my choice. But most likely I will stick to Aegis.

Yes, this clear text export is a potential security risk. I get it. But if it means I have a way to easily migrate 30+ 2FA accounts I'm willing to make that compromise. Yes, I mean.. Now that I have all my secrets and otpauth:// URLs that shouldn't be a concern anymore, right? Well, now I have everything. I'm pretty sure in the future I'm forgetting to properly document some 2FA codes again, hence this being the better choice.

Or are there other solutions I'm missing?

And what about the Microsoft Authenticator?

Honestly? I'm forced to use this by my employer as we don't allow any other form of 2FA for authentication in our company. This is a configured setting in our MS365 organisation. Therefore I can't replace the Microsoft Authenticator with another app.

Additionally the Microsoft Authenticator also implements the so-called "Number matching" 2FA method (this blogpost provides a bit more detail: Advanced Microsoft Authenticator security features are now generally available!). This is where the user has to enter a 2 digit long number shown on another user-owned device. It was developed to combat the Multi-factor authentication fatigue where users are bombarded with dozens or hundreds of authorisation requests - until they finally click accept and allow the attack to be authenticated.

And while it seems that Google also implemented this method, it seems that the technical background workings are strictly coupled with the Microsoft Authenticator app. And it seems that neither Aegis nor FreeOTP support this method yet.

Hence there is only one account tied to it. So I did what was reasonable: I removed the app, deleted all settings and cached files, reinstalled the app and just enrolled my account again.

Yes, this required a ticket for our IT Helpdesk to remove the old Authenticator from my account, but I had no problem with that.

Photo by Pixabay: https://www.pexels.com/photo/red-pencil-on-top-of-white-window-envelope-236713/

I had already forgotten that Mozilla bought K9-Mail in June 2022 in order to transform K9-Mail into Thunderbird on Android. Now I was reminded again as on October 30th 2024 the first Android version of Thunderbird was released.

However the initial beta releases were accompanied by a little bit of drama regarding the data privacy topic. As the first releases of the Thunderbird App contained telemetry trackers from Mozilla and those were enabled by default (Opt-Out instead of the more data privacy friendly Opt-In). Additionally the user wasn't made aware of this during the install and configuration process.

These facts became aware to many users through the following GitHub Issue: Thunderbird Issue 8199: Expose the ability to mange Telemetry settings on first-time use where the reporter just stated in a factual way that he expects these settings to be off initially.

However the first reply to that issue didn't make things better. Apparently a Senior Manager/Mobile Engineering at MZLA Technologies Corporation, the subsidiary of the Mozilla Corporation of which Thunderbird is now a part of, wrote the following as a reply:

Unfortunately we cannot make this type of data collection opt-in because the limited data from voluntary reports wouldn’t provide enough insights to make informed product decisions. Opt-in data would come from a small, biased subset, leading to flawed conclusions.

Knowing the Android ecosystem covers a vast range of hardware and form factors, we need to have a mechanism to make better decisions on how features are being used, and have information in which environments user might be having trouble.

In line with Mozilla’s data practices, the default data collected contains no personal information. This helps us understand how features are used and where issues may occur, while minimizing data points and retaining only what's necessary. When we decide on new probes, we actively consider if we really need the information, and if there are ways we could reduce the needed retention time or scope.

While I can't offer an opt-in at this time, I understand your concerns and genuinely appreciate that you're thinking critically about privacy. You might also be interested in a recent talk about our need for privacy respecting telemetry. https://blog.thunderbird.net/2024/08/thunderbird-goes-to-guadec-2024/

This again sparked a lot of comments who can be sorted into the following categories:

1. Disappointment that an application developed by Mozilla uses such shady practises. Along with criticism that users are not informed about this and there are no information on what type of information is gathered and how it is used.
2. Notices on the various laws forbidding such data collection (especially the GDPR from the EU).
3. Sadness that while K9-Mail was tracker free, Thunderbird obviously won't. Which disappoints many data privacy focused users.

Or as someone, sarcastically, pointed out on Mastodon (Source):

How could K-9 be developed and become the best email app for Android, and even make ‘informed product decisions’ without a tracker? Sarcasm over.

With the 8.0b2 release that feature was removed and will, hopefully, be reworked in a more user-consenting way.

Personally I am also very disappointed and my anticipation has taken a huge blow. Mozilla once stood as a beacon of user-centred interests. And while I wholeheartedly agree that they should be able to get usage metrics I too want this to happen in an open and consenting way. Enabling the user to actually make a choice and inform me about the nature of the data being transmitted.

Other resources

There is an FAQ what will happen to K9-Mail and Thunderbird in the future: https://blog.thunderbird.net/2022/06/faq-thunderbird-mobile-and-k-9-mail/

The roadmap can be found here: https://developer.thunderbird.net/planning/android-roadmap

Photo by cottonbro studio: https://www.pexels.com/photo/man-observing-woman-through-doorway-8626372/

When Apple introduced the AirTag (see Wikipedia) it was primarily marketed as a "Find your device/stuff" product. Allowing you to locate the item to which the AirTag is attached, even when it's hundreds of meters (or kilometers) away. As long as there is some Apple device which receives the Bluetooth signal and forwards this to the Apple servers, you will have the location of the device. Of course, depending on time passed and location it can be inaccurate. But in our connected world it's likely that some device will pick up the signal again and you'll have up-to-date location information.

Additionally AirTags can produce a sound, so that you can get a audio hint on where the device is.

AirTags do have many useful cases. From tracking your stolen bike (if you hide some AirTag on/in it), to locating your lost luggage at an airport.. Even pets! Sure this is useful. But sadly.. The principle of dual usability is real and hence even in the beta phase Apple already rolled out a feature that allowed you to view all AirTags in your vicinity. As the potential for illegitimate usages was too high, to simply ignore it. After all.. Watch someone retrieving money at an ATM, occasionally bump into this person and put an AirTag into the jacked of that person. And then just follow and wait until this person is in some place where there are no cameras and/or eyewitnesses. Or think about the whole stalking and Online-Dating problem.

No, AirTags can be a security and privacy disaster. I'm sure many of you have read the story about Lilith Wittmann, who used an Apple AirTag to uncover an office of a secret german intelligence agency (article on appleinsider.com) or read the original german article, published by herself, here: Bundesservice Telekommunikation — enttarnt: Dieser Geheimdienst steckt dahinter.

Ok, so Apple has this feature included in iOS for it's phones/tablets, etc. - What about Android?

Good things first: Apple and Google recognized the threat and are working together towards an industry specification which aims to put an end to stalking via AirTag and similar devices. But, as this has just been announced in May 2023, it's still too early to have produced any meaningful results (sadly).

Apple press release: https://www.apple.com/newsroom/2023/05/apple-google-partner-on-an-industry-specification-to-address-unwanted-tracking/
Google blog post in their security blog: https://security.googleblog.com/2023/05/google-and-apple-lead-initiative-for.html 

Well, Android being the fragmented Android market it is, not every manufacturer has such an option included. I know that Google Pixel devices have such a feature. And I was told Samsung and OnePlus phones too. But there are many other Android versions around. And: What about custom ROMs? I use LineageOS on my OnePlus phone and wasn't able to find such a feature. That's why I searched for an app that does this for me, and was pleasantly surprised to find one.

AirGuard is even released for iOS and there it's also able to find trackers which Apples feature won't detect. So.. I guess this is a recommendation to install this app on iOS too.

Introducing AirGuard

AirGuard is an Android app, developed by the Secure Mobile Networking Lab (SEEMOO) which is part of the Technical University of Darmstadt - specifically their computer science department. You may have heard from them occasionally as they regularly find security vulnerabilities in Apple products and do a lot of research on Bluetooth and Bluetooth security. The neat point? AirGuard is OpenSource, it's code is being published on GitHub. This allows me to install the App using F-Droid (which only offers OpenSource apps).

The icing on the cake? It can not only track Apple AirTags, but Samsung SmartTags and Chipolo Tags too.

From here it's just a normal app installation. Allow the app to use Bluetooth, disable battery saving mechanisms (so it stays active while being executed in the background) and that's it.

As I own no AirTag or similar device I can't test it, but I will update this article when I was able to test this.

If you want to stay up-to-date with the development, there is a Twitter account for that: https://twitter.com/AirGuardAndroid

Photo by Pixabay: https://www.pexels.com/photo/black-android-smartphone-on-top-of-white-book-39584/

Too Long;Didn't Read (TL;DR):

You can omit the steps listed below. If your 2FA/OTP App allows to specify the secret key, type, algorithm and interval use the following settings for LinkedIn:

Type: TOTP
Number of digits: 6
Algorithm: SHA1
Interval: 30 seconds

Original article

I try to enable Two-Factor-Authentication, or 2FA in short, on any of my accounts that supports it. But: I dislike it, when the 2FA-Codes are sent via Mail or SMS. This is just too insecure as both can be intercepted. And personally I would go so far to say "SMS & Mail isn't a valid & secure second factor." As there are too many reports how scammers and phishers intercept SMS or mails. Yet many companies still default to this. LinkedIn too.

Therefore I wanted to switch to my Authenticator App of choice: FreeOTP - https://freeotp.github.io/
The source code is on GitHub: https://github.com/freeotp

It is completely OpenSource (sponsored by RedHat) and even available in the alternative Android App-Store F-Droid, which only offers Apps which can be build completely from source.

As naive as I am sometimes I thought it's just the following steps:

1. Enable 2FA in my LinkedIn profile
2. Provide password to authenticate
3. Scan the QR-Code in FreeOTP
4. Enter the generated code to verify it works
5. Generate & Save the backup keys in my password manager

But not so on LinkedIn. They don't display a QR-Code. Well.. To be precise. They did. Before Microsoft bought LinkedIn. After that this changed. Nowadays they only display you the so-called secret key (encoded in Base32) and that's it.
Then LinkedIn tells you to install the Microsoft Authenticator App, while mentioning, that you can, of course, use any other Authenticator App.

The problem? The described workflow on what to do with that key only works in the Microsoft Authenticator App.
Side-Note: Someone told me Google Authenticator should be able to use that code too. But I can't verify this.

LinkedIn gives you absolutely no additional technical information.

• No otpauth:// URL
• No information if TOTP or HOTP must be used
  • Well, to be fair, we can safely assume it's TOTP.
• Which algorithm must be used?
• What is the lifetime (interval) of the generated codes?

Nothing. But this is what I need with FreeOTP. I tried a few combinations, but had no luck.

So I resorted to the Linux command-line.

1. Enable 2FA in your account until the secret key is displayed
2. Install qrencode (or use one of the available Web-Generators for QR-Codes at your own risk)
   • Source Code is here: https://github.com/fukuchi/libqrencode but every major Linux distribution should have a package available
3. Build the following string: otpauth://totp/LinkedIn:MyAccount?secret=KEY-YOU-GOT-FROM-LINKEDIN
   • All in one line, no spaces at the end, no enter.
   • You can change "MyAccount" to something more meaningful like your mail address
   • Example: otpauth://totp/LinkedIn:JohnDoe@company.tld?secret=U4NGHXFW6C3CLHWLQEVCBDLM5FQMAQ7E
4. Paste that string into a textfile.
   • Again, no enter or spaces at the end
5. Execute: qrencode -r /path/to/file.txt -t png -o /path/to/image.png
   • This will generate a PNG-Image at the location specified by the -o parameter
   • -r is the input file containing the string
6. Display the QR-Code and scan it with FreeOTP
7. Verify the code works
8. Generate your backup keys and save them in your password manager
9. Profit!

Some documentation regarding the otpauth:// URL, it's syntax and the parameters you can use is available in the old Google Authenticator repository on GitHub: https://github.com/google/google-authenticator/wiki/Key-Uri-Format
(Google Authenticator once was OpenSource too, but sadly isn't any more.)

And while at it, I created the corresponding GitHub Issue for the FreeOTP project #360 [Feature-Request] Allow adding of entries by just specifying label & secret to properly take care of this nuisance. ;-)

Lessons learned

FreeOTP assumes the algorithm of SHA1 and an interval of 30 when these parameters are not part of the otpauth-URL. Choosing these works out of the box and I can omit the QR-Code step this way.