old people and security

Let's say I am in contact with a bunch of people that are not really tech-friendly, but that are in urgent need of secure means of communication. I have to teach them how to protect themselve mainly from government institutions. The problem is that I have little idea where to start. I was first thinking about riseup.net, vpn and pgp encryption, but it may be a bit too much. Their idea of safe communication is meeting in a no-phone room every week for 2 hours, I was hoping I could find something less old-school and a bit more practical. Any advice? How would you proceed?

https://protonmail.com set them up all with this and if possible a cheap laptop like a chromebook running TAILS os https://tails.boum.org/

>>3757 Are they on the phone? Signal? Telegram? >>3758 This

Are burner phones still practical? a lot of places now require ID and a contract to purchase phones.

fuarrrk protonmail, fuarrrk telegram, fuarrrk whatsapp, even fuarrrk signal. fuarrrk all that noise if its gov threat model. whatsapp and signal groupchat both got compromised like last thursday because of the stupid group id's scheme for giving out keys -- and more sick vulns to come surely. maybe even on the main 1-to-1. Tell them to avoid communicating through technology all together and if they have to use it, if there is no way around it, teach them pgp and coordinate a way to key exchange once a month. just only teach them pgp... can be used literally anywhere for end-to-end encryption. can be used to protect files too. decoding and encoding from a main pgp application will cause them to think about what they want to say and consider whether they want to send it far more. let them use gmail fi they fuarrrking want It's not like their messages always have to be encrypted right? so dont... dont paint a giant target on their heads. other than this, what more could they need? maybe an uncompromised system... so teach them how to update their computers and phones and if you have time buy them processors with good hardware virtualization support and teach them to run virtualized enviros. this is a silly scenario op

>>3763 p.s. i say fuarrrk protonmail because easy phish and relies on https stupid ca model for entry of both passwords. in any oppressive regime your users are gonna get shoah'd

>>3757 They may not be tech friendly, but you're asking for advice on a technical (ish) board, so I'll respond with technical knowledge, especially as it is likely relevant given the nature of our cyberpunk dystopian world today. Firstly, what exactly is their threat model here? Is this passive (ie, someone might find out later, or overhear something) or active (someone actively, and directly, investigating) threat? Also government could cover many governments, and we would need to adapt some methods for certain governments, and institution could be anything from the DMV to Mossad. In the latter case, there's not really anything you can do; you gon get Mossaded. I'll send flowers or something. Next, a no-phone room doesn't do much good if the phones are just left outside the room or something, they need to be left home, or placed in a microwave (or fridge) somewhere else. Phone locations can also be triangulated, so if they're trying to avoid being associated with each other and their particular threat model has legal power in their country to compel a telecom to divulge CDR and location records, that's bad. However, even if it's oldschool, it's not a terrible method, and it's a lot better than a poorly implemented technical solution. Depending on the country, I will say the first step is probably going to be Tor. Tor is significantly better than a VPN, UNLESS we are dealing with a country where it may be a crime to use Tor, or will cause significant problems down the line (Iran, China, etc). This may be through Tails or another bootable ISO or something. We want to use Tor instead of a VPN because a VPN is single-hop, lots of VPN providers are fuarrrking godawful about security, and again, they can be legally compelled (I know they say they don't keep records, but use your fuarrrking head). In addition, if they all use the same VPN and their government institution opponent learns about it, they'll just start associating people using this VPN and use that to build a case. In terms of instant messaging, I'd recommend either a Jabber/XMPP client with unique addresses for each person they need to contact, connecting through using busy servers and using OTR (or OMEMO, depends on how much you like Moxie or whether you need multi-end to multi-end/offline message crypto). Ricochet is another reasonable solution. PGP encryption is a fine idea, but it's very easy to fuarrrk up, it leaves lots of metadata, and it's generally quite complex. I actually use PGP almost daily, and I do not enjoy it in the slightest. Also, email isn't great. Basically, this is a hard problem. In terms of threats: if these people are being actively targeted by whatever their government opponent is (let's say, for example's sake, that it's a repressive regime investigating journalists), they cannot do this from their house or connection associated with them. It's very plausible that the ISP will have a lawful intercept tap placed downstream, and traffic correlations may be enough to get them all fuarrrked on conspiracy to defame the glorious people's dictator or whatever charges. You cannot identify whether this has been done, before you say anything about trying to find out. There's a strong likelihood they are being physically monitored. That is not something you can really do anything about, and in all honesty, I'm not an expert about physical surveillance (and it's unlikely anyone here is). In addition, you'll also need to realise that some governments make use of malware themselves. These people should only be accessing their normal emails from within some kind of virtual machine, and they should save a known good state before doing anything, and revert to that after doing anything. As a word of warning, if this regime business or something similar is the case, you probably shouldn't be getting involved. I know you post on lainchain, but the fact you had to ask means you are not qualified and it might put you at serious risk. I hope you don't take offence at this, it's just an unfortunate point that needs to be made. If it's just old people fuarrrking around with THE BLOCKCHAIN or whatever, go nuts.

>>3763 I strongly agree with the "don't use groupchat" comments; I'm unfamiliar of any actual compromise with double ratchet, but basically, just don't. It's not good for EVERYONE to have the ability to have logs of EVERYONE ELSE.

>>3765 Oh, lastly, while I really love this imageboard, just be aware that if this is a real scenario, and people really are potentially in danger from a repressive regime, there are a whole lot of hacker LARPer-types here. Even me, on my best day, doing this for technical people would likely only be able to go as far as protecting them from passive stuff, and there's still the possibility someone would fuarrrk up. Going toe-to-toe with an active intelligence or law enforcement agency is pretty much suicidal. There are a lot of better organisations out there, such as Privacy International or Citizenlab, who are equipped to deal specifically with journalists in repressive countries and provide support. But again, if this is just old fuarrrks talking about BLOCKCHAIN, or Elderly Anarchists Meetup Group With Coffee and Biscuits, things are less of a concern.

These are what I've got, it's probably a bit old by now.

>>3758 >>3764 Protonmail could be a solution, altough not the most secure, I do not really know how far law enforcement will go in order to compromise their communications. However I will probably have to make a dozens of TAILS live usb. Cheap and old laptop shouldnt be a problem. >>3759 Even though this could be the easiest way, most of them do not own a phone capable of running these apps, so it is not really a good solution. >>3760 I have though about burners but as you wrote it requires an ID and a contract. >>3763 I should be able to teach them how to update their devices, but pgp will probably be more tricky. But I'll talk to them about it. Virtualized environments are also doable since the complex part is more in the setup. >>3765 Threat model here is currently passive (I hope), but could become active in the next couple of months, depending on the evolution of their situation. (I don't really know all their plans) Tor and TAILS could really come in handy, as it is not illegal yet. As for Jabber/XMPP, would you recommend a client in particular? I have heard of Ricochet but is it reliable? their github page states "This software is an experiment.[...] Do not rely on Ricochet for your safety". Concerning physical surveillance, search or raid, there is nothing I can do much about, they probably have more experience about it than I will ever have. If I ever get physically monitored I'll be fuarrrked. I am aware I am not qualified enough to correctly face this problem, and I would rather let them stay old school than implement a flawed technical solution, but they are willing to learn to communicate more frequently and securely between meetings and to be able to reach each other in case of an emmergency. So I though I'll help them, no one is going to do it otherwise. >>3769 >>3770 Thanks I'll take a look! Hopefully no one is going to get mossaded or anything else, as I said I don't really know how many ressources will be used against my old fashionned political dissidents. Anyway thanks for the solid advices in this thread, I have a better understanding of what to do now.

>>3778 I'd recommend something like https://github.com/agl/xmpp-client, or maybe https://coy.im/ As far as being reliable, no software is... reliable. It is, however, if I recall, easier to use than XMPP, so that might be good. And I don't believe it has any glaring flaws, though I may have missed something in the news recently.

>>3765 What's wrong with PGP's metadata if you don't use some identity server or sign your messages, just store a little text file for your contacts: Bob - KEY1 Alice - KEY2 John - KEY3 And I want to add about Tor, if it's somehow blocked in a country, it is possible to use obfuscated bridges, yes they are not a perfect solution, but they are designed to bypass firewalls, not to protect against whitelists, database bruteforce or targeted surveilance that corellates user's traffic on honeypot hidden services. If Tor is blocked and people are afraid to even use bridges because they don't want to be asssociated with it, then shadowsocks, stunnel or similar obfuscating protocols for first hop are a viable solution. I guess, if you put it on some roaming-IP VPS like amazon's, they would be completely off-limits for 3rd world dictatorship's DPI monitors and hackers.

>>3778 As far as I know, Protonmail does give up connection logs for users upon request from foreign legal entities, blocks user accounts and is under jurisdiction of Switzerland law, which forces companies to keep logs since last year. Encrypting email locally and sending it over any email server is simply more secure. Do not use burners, phone network is compromised, it can be tapped locally or remotely, called ID can be spoofed, voice fingerprints are collected here and there, etc. Do not talk or send text messages unless it's literally one-time burner use. Unless you know what you're doing and use cellular internet capabilities only, it is more like using your home fiber connection but with more metadata (alternating location). If you are in a big city, consider street surveillance cameras and private CCTV a viable threat, today Stasi-like physical surveillance is not needed if a target lives in a city like London with more cameras than people. Think of storing valuable data in encrypted filesystem containers on microsd cards, they can be easily destroyed. If data is too big and requires conventional HDD, store salted keys on microsd cards, if key is gone, data is irretrievable. Boot an OS from microsd cards of flash drives (bootloader and key for root partition) and let your users always carry those with them to prevent simple case of Evil Maid attack on their hardware when a disk bootloader is altered by malware (advanced tampering still can't be proved).

>>3778 on protonmail entire mailbox is chilling decrypted in memory rather than per email decryption, do you even control your private keys in it? because I don't remember being able to upon testing. session data is always a thing too using elliptical curve cryptography which is fragile / could be backdoored (see Daniel J Bernstein and bruce schneier papers on subject) inside tls... probably 1.2... and while we're on that subject who knows what TLS downgrade policy your users have... or what certs are already in their trust store... probably bluecoat at least right? yep. not to mention those invisible popups that steal from your browserpassword autofiller / formfiller its just not a good option unless they are using it for one email message then baleting or something. there is far more to fuarrrk up by depending on 90 different applications with implementations you don't know rather than just teaching them a valuable skill like pgp and then focusing the entirety of your effort on endpoint security, which is super weak always.

>>3782 >If you are in a big city, consider street surveillance cameras and private CCTV a viable threat I'm lucky they were illegal in my country. At least until they figure a way around the legalities. >>3784 >do you even control your private keys in it? because I don't remember being able to upon testing. You can (or could, last time I checked) but everything else you say is valid. to be honest people should just use protonmail as another throwaway.

>>3779 agl is a genius but otr is mostly focused on synchronous communication resulting in more re-verification, to which a bunch of uneducated dissidents are probably just going to click verify without verifying fingerprints using an alternate channel. omemo builds upon otr and is much better at asynchronous messaging so slightly less risk (less re-verification). conversations (free and open source on f-droid) allows omemo and pgp public key cryptography . agree on ease of use not being xmpps strongsuit. decentralized chat protocols tend to have large overhead and in general xmpp is kind of a mess every time i've used it.

>>3760 >>3778 >Are burner phones still practical? a lot of places now require ID and a contract to purchase phones. >I have though about burners but as you wrote it requires an ID and a contract. I can't speak for everywhere, but it's really, really easy to get a burner phone here (USA). Just grab any of the pay-as-you-go phones and some airtime cards for them at Walmart. There's even a display of them at my grocery store. I use one of the more popular brands (TracFone) for my cell phone. Over the past 8 years or so, I've bought several phones under the same brand. I usually order new phones online, but a few I bought in-person at the store in cash and was never asked for an ID. You do have to register the phone online (unless your transferring your minutes to a new phone from a previously registered one) and they require your name, address, email, and such, but there's no verification of your identity, and you can easily lie. They don't ask for anything sensitive that they could use to verify your info like drivers license number or SSN, either. But like >>3782 said, if the means of communication is meant to be secure, burner phones aren't a good idea. Burner phones are more of a "security through obscurity" thing where malicious actors can't attack you since they don't know your phone number, but you lose that obscurity after you use the phone.

I'd set them up with persistent Tails USBs with at least one of either: IRC+OTR, or EMail + PGP. Then just explain very clearly to not connect to these channels via anything other than their Tails USB. You could then set them up with a mailing list to post to. In doing this you would need to help them to set up Enigmail in Thunderbird to encrypt messages sent to the mailing list for the entire affinity group.

>>3788 Just buy it from some sketchy guy with a stand or something

>>3788 to add to your points: i think laws around burner phones might vary state-to-state, but i don't know for sure. they're easy to get in every state i've lived in. also regarding security, the benefits of the burner are only fully realized if you actually regularly burn it. it's not exactly practical to do that if you're using the phone as your main contact number. burner phones are probably the lesser evil vs. a cell company contract with your real identity, but you are right to caution against a false sense of security. OT: what about SIP services?

>>3757 I would send the to Florida

>>3788 I think that states that are anal about restricting phones and sims basically just don't have their soykaf together when it comes to digital surveillance. The fact that Western countries don't think twice about letting anyone an anonymous sim or phone tells you all you need to know about the security and privacy of the cellular system.