Hacker News new | ask | show | jobs
by cookiengineer 1064 days ago
Use briar [1] for online+offline end-to-end encrypted messaging. Use LineageOS [2] without gapps (aka without Google Play Services) and get a device that is officially supported with current LineageOS version. Don't use any XDA developer builds, because they're known to be infected with malware.

Obviously don't use Meta or Google apps, because that's where the backdoors are for governments. Don't use WhatsApp, don't use Telegram, don't use Threema. They're compromised.

Use AppWarden [3] to enable/disable/verify the usage of known trackers in your apps.

Use NetGuard [4] as an Android firewall.

Use F-Droid [5] and Fennec builds [6], with uBlock Origin to protect your smartphone from malvertisements.

Never synchronize your contacts, block contacts access for all Apps; and make sure you don't use their real names. Contacts stored on or accessed by SIM cards (e.g. call history) can be downloaded via Class 0 SMS, remotely.

If possible, I'd avoid MediaTek based SoCs because their rootkit was leaked a couple years ago and it works still on newer chipsets. I would recommend an "as open source as possible" device, like the Google Pixel devices or the Fairphones.

On your Desktop or Laptop machines you should switch to a Linux distro of your choice. The most reasonable secure ones are Arch (not beginner friendly), Manjaro, OpenSUSE - or as a beginner friendly alternative - LinuxMint.

Would advise against Debian/Ubuntu though for security reasons (which would include LinuxMint).

The Arch maintainers (and therefore Manjaro, too) heavily reduced the attack surface of SUID binaries or LOL binaries that could be abused for privilege escalations and/or remote exploits/persistence etc. [7]

[1] https://briarproject.org/

[2] https://wiki.lineageos.org/devices

[3] https://gitlab.com/AuroraOSS/AppWarden

[4] https://netguard.me

[5] https://f-droid.org/

[6] https://f-droid.org/en/packages/org.mozilla.fennec_fdroid/

[7] https://gtfobins.github.io/

edit: clarification of LinuxMint

2 comments

I really don't want to derail the thread as it is useful, but i'm curious about the security reasons against Debian/Ubuntu.
I'm also curious how Mint resolves the security threat of Debian and Ubuntu, given its lineage.
For reference I've included the OVAL datasets [1] [2] for Debian and Ubuntu, as well as the CVE list [3] for Debian that's used as a basis for generating the issues/view in the frontend. For argument's sake (feature freezes are bad for security is my baseline argument) I'm focussing on Ubuntu LTS which is jammy at the time.

The issue terminology you want to look for is "diverged too much from upstream" or a bunch of different similar EOL or end-of-life or "end of life" tags that are not standardized in any security tracker's format. I'd argue that both security teams use a software that uses free-form textareas for setting the tags of the issues (e.g. closed, fixed, and other).

Accumulative word list that might be incomplete, to discover those kind of issues:

- "bug, not a security problem"

- "cannot reproduce"

- "eol"

- "end-of-life"

- "end of standard support"

- "end of esm support"

- "out of standard support"

- "reached end of life"

- "ignored"

- "changes too intrusive"

- "contains no code"

- "code is different"

- "code is very different"

- "code not built"

- "code not compiled"

- "code-not-compiled"

- "code not present"

- "code not-present"

- "code-not-present"

- "code not shipped"

- "not-in-code"

- "disputed"

- "fix would break"

- "intrusive"

- "not remotely exploitable"

- "was deferred"

- "was needed"

- "was pending"

- "no server code"

- "no update available"

- "no security impact"

- "not available"

- "not security vulnerability"

- "not supported"

- "not upstream fix"

- "ugly backport"

- "update not available"

- "upstream version is not redistributable"

- "removed from archive"

- "replaced by"

- "superseded"

- "superseded by"

- "too intrusive"

Source: Am building a scraper and vulnerability database that is cross-distro, and has different confidence factors for different linux distributions (for mentioned reasons). [4]

[1] https://www.debian.org/security/oval/oval-definitions-bullse...

[2] https://security-metadata.canonical.com/oval/com.ubuntu.jamm...

[3] https://salsa.debian.org/security-tracker-team/security-trac...

[4] https://github.com/tholian-network/vulnerabilities

Why would Mint, a desktop-oriented fork of Ubuntu, be more secure than Ubuntu?
No clue if they are right. If they are, I can imagine it is because Mint would be stripped of any telemetry and other corporate bits of Ubuntu that Canonical has in there. From what I know, Canonical/Ubuntu does a good job of anonymizing that data and doesn't capture anything sensitive and put it over the wire, however in the worst case scenario of a coup, it probably doesn't hurt to be over cautious.

Although, assuming worst case. I personally say, OpenBSD.