Cryptography nerd

Fediverse accounts;
@Natanael@slrpnk.net (main)
@Natanael@infosec.pub
@Natanael@lemmy.zip

@Natanael_L@mastodon.social

Bluesky: natanael.bsky.social

  • 1 Post
  • 656 Comments
Joined 3 years ago
cake
Cake day: August 16th, 2023

help-circle
  • Depends on how repressive.

    Often your main method of staying safe is appearing harmless.

    You can find dedicated Mastodon and lemmy hosts more receptive to VPN users.

    Anonymous use is hard due to stuff like timing attacks and writing style recognition, etc, especially if you post publicly. You want to mimic another style, for example. And create plausible deniability around timings (like say scheduled posts)















  • For sites you visit occasionally, it’s better to enable tab isolation (use the containers feature) and then enable JS only for that domain (note the difference between allowing JS from that domain in any tab, vs only allowing that tab with that domain to use JS, you should do the latter)

    https://addons.mozilla.org/en-US/firefox/addon/multi-account-containers/

    If you’re switching to a different browser you may as well use the same browser but a second clean profile and use private tabs so it doesn’t retain history. Using private tabs in your main browser profile does also help but isn’t perfect because there’s still some metadata leaks occasionally.

    Using a different browser could ironically make you easier to track - how unique you are is the main signal used to track you (user agent, OS, language, etc), and going for an even more rare config will help their tracking even if you delete session cookies. Especially if they have a tracker across multiple domains you visit from different browsers from the same IP, with similar device fingerprinting results across browsers. That’s a strong signal those sessions are linked. You want to NOT stand out to maintain your privacy.




  • There’s also a big difference between published specifications and threat models for the encryption which professionals can investigate in the code delivered to users, versus no published security information at all with pure reverse engineering as the only option

    Apple at least has public specifications. Experts can dig into it and compare against the specs, which is far easier than digging into that kind of code blindly. The spec describes what it does when and why, so you don’t have to figure that out through reverse engineering, instead you can focus on looking for discrepancies

    Proper open source with deterministic builds would be even better, but we aren’t getting that out of Apple. Specs is the next best thing.

    BTW, plugging our cryptography community: !crypto@infosec.pub