Packet sniffer? Like wireshark maybe?
this post was submitted on 04 Oct 2023
79 points (98.8% liked)
Firefox
17794 readers
29 users here now
A place to discuss the news and latest developments on the open-source browser Firefox
founded 4 years ago
MODERATORS
But wouldn't that just tell you, "Firefox was connecting to the internet"?
Nope, it'd give details. Combing through and decrypting those details is another matter, though.
You can MITM yourself and setup your own CA, no?
If you so chose to, yep.
No, it would say which packet is being sent to which destination.
It just isn’t easy. You can read permissions and try to read the source, but it requires some expertise.
For smaller extensions, you can start by looking at the manifest.json as this file is the entry point. You can see which part of the code needs access to 'everything' and go from there.
https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/manifest.json
For example, with Instance Assistant for Lemmy, you need to give access to all sites because it needs to run on any Lemmy site and that's not something you can hardcode in.
I think it's "access your data" permissions that are the ones to be wary of, due to the explanation here. Defending against this, I'm not really sure. Someone who knows more should chime in, but maybe a software firewall like Little Snitch/OpenSnitch that will let you approve/deny every connection. (This will probably get fatiguing fast.)
As a longtime Little Snitch user, it's freakin exhausting.
I thought little snitch worked per app and not for each connection one app makes
You can make rules network-wide, per-app, or per-incident. The latter is useful for getting a handle on app behavior. Like if you see it contacting 'updates.somedev.com' weekly, you can choose to allow or disallow permanently based on how benign you think the app is. But more likely, anything trying to phone home has a dozen CDNs it's trying to hit rather than an easily identifiable URL. Block one, it tries to hit the other. Maybe today, maybe next week. It gets overwhelming (which IMO is a feature for the dev, not a bug).
Thats cool
A. PiHole has the added benefit of blocking some known blacklisted domains and even if it doesn't block it, you'll see the DNS resolution requests by device in your network. It might be hard to know which requests are coming from the add-on if the hostname or timing is not obvious though.
B. Using a local sniffer like Wireshark you can list everything that goes through a network interface in real time. Since it's a really long list, I'd isolate the app first by running it in a container and only sniff that container's interface instead. And since you're using containers, you can also run Wireshark itself in one. This is an example of Firefox in Docker - though I've never used it.
C. If open source and you have the knowledge, check the source code for mentions of request protocols "http", "tcp", or keywords like "timeout", "payload", "track", and "telemetry". It's common for projects to use libraries that collect and submit logs, often for legitimate purposes, like sentry, new relic, and datadog.
Thanks! I haven't yet tried to use wireshark with a container, tho'll keep that in mind, but I did check the code on Github (my coding knowledge is small, but not nonexistent, and it was easier once I had an idea of what to look for). The addon I was interested but wary of at the moment was Redirector, but from what I've seen so far it seems legit (I didn't find any occurence of most of these keywords, or anything that seemed suspicious with the occurences of http).
Easiest way is to use packet sniffing. but they could be set-up to only send data at random intervals, which could be hours, days or weeks. so you're best bet is go through the source code by extracting the extension file and look for any links, domains or mangled code which may be encoded with something malicious.