SimpleX Chat

SimpleX - the first messaging platform without user identifiers of any kind - 100% private by design!

Security assessment: https://simplex.chat/blog/20221108-simplex-chat-v4.2-security-audit-new-website.html

website: https://simplex.chat
github: https://github.com/simplex-chat/simplex-chat#readme
donate: https://github.com/simplex-chat/simplex-chat#help-us-with-donations

Public Key

npub1gsunl7x2y8ka4rn72y9zmzh0j8ntw8ln3956dxe4vg2mldpt4svsuzfy6l

NIP-05 Address

[email protected]

Profile Code

nprofile1qqsygwfllr9zrmw63el9zz3d3there4hrlecj6dxnv6ky9dlks46cxgpzpmhxue69uhkummnw3ezuamfdejsj236xq

Publishing to

nostr.wine

Author Public Key

npub1gsunl7x2y8ka4rn72y9zmzh0j8ntw8ln3956dxe4vg2mldpt4svsuzfy6l

Show more details

Last Notes

2024-07-11 00:02:39

- reply

I don't think it would have a major difference as the app doesn't check messages, it is push. But if you do an experiment, please let me know. Using private routing should reduce battery usage a bit.

2024-06-20 20:04:38

A big win for the common sense!
All politicians and supporters of mass surveillance for the sake of "protecting children", whether they are corrupted or genuinely misguided, should realize that #ChatControl approach does not make any sense.
If personal data of families and children ever becomes available to any third parties or algorithms, it will be leaked and used by criminals to exploit and abuse children and vulnerable people.
We wrote about it previously here: https://simplex.chat/blog/20240601-protecting-children-safety-requires-e2e-encryption.html

2024-06-04 21:38:13

- reply

updated!

2024-06-04 21:29:19

SimpleX network: v5.8 released with private message routing, IP address protection for messages, files & media, new chat themes and more!
Also new in v5.8:
- group improvements - reduced traffic and additional preferences.
- improved networking, message and file delivery.
- Persian interface language - thanks to our users.
Read more: https://simplex.chat/blog/20240604-simplex-chat-v5.8-private-message-routing-chat-themes.html
Install the apps: https://simplex.chat/downloads/
Please upgrade your self-hosted servers!
See the new server guide: https://simplex.chat/docs/server.html

2024-06-01 22:53:49

Protecting children's safety requires privacy and end-to-end encryption:
https://simplex.chat/blog/20240601-protecting-children-safety-requires-e2e-encryption.html
Proposed "upload moderation" would fuel the very problem it aims to solve, undermining rather than protecting children's safety.
#safety #privacy #encryption

2024-05-28 20:43:05

Please sign the petition to the EU Council by the Global Encryption Coalition about the proposed "upload moderation" that would create serious security and privacy risks:
https://actionnetwork.org/petitions/global-encryption-coalition-joint-statement-on-the-dangers-of-the-may-2024-council-of-the-eu-compromise-proposal-on-eu-csam
We all want to solve child abuse problem, but these proposed measures will make it worse:
- centralized scanning of family photos risks leaking them to criminals, increasing child abuse.
- the criminal case against Meta in the state of New Mexico shows how its content algorithms enable child abuse.

2024-05-17 18:50:26

- reply

right:) 👍

2024-05-17 18:49:49

- reply

I agree about full disclosure once fixed - it's what our policy is, similar to OpenSSL's.

2024-05-17 18:47:13

- reply

two more assessments are coming this year - one focussed specifically on the protocols design, and another on implementation.

2024-05-17 18:44:57

- reply

In DR neither party has control of the AES key to use, so maybe I am missing something, but I don't think this attack works... will re-assess.

2024-05-17 18:43:30

- reply

no, we absolutely love criticism - it's what was driving the design evolution so far, and will continue.
Just please send any real findings (if you get deeper) privately - as real people security depends on the app, whether you like it or not.
But the criticism of the design itself is very appreciated.
So far it's all good. AES-GCM given how it's used seems adequate, but we will revalidate it.

2024-05-17 18:39:38

- reply

Whether you like design or not is secondary to whether it is a secure or not, isn't it?
Design evolves though.

2024-05-17 18:37:36

- reply

arithmetics says it's 4 :)
And they generously billed 5 days having spent more than 10 (engineer-days).
TOB are very good with Haskell btw, and double ratchet implementation was one of the focus points.

2024-05-17 18:35:09

- reply

> My complaint with Curve448 is that most things don't use it, so implementations are less reviewed.
OpenSSL (both library and the tool) includes it for a long time.
> The primary reason to select 448 over 25519 is a "biggest rock is best rock" approach to cryptography parameters. It doesn't help in any realistic scenario, nor against quantum attackers at all.
This view is contrary to the advice from DJB to always choose a bigger key size that is necessary.

2024-05-17 18:33:26

- reply

we understand the limitations of AES-GCM, and they do not apply here, but we will revalidate it.
Groups don't use shared keys, they are based on pairwise ratchets.
The encryption scheme evolved with the addition of sntrup761 to double ratchet, and if AES-GCM proves suboptimal, it can be replaced too.
Also, it is customary to follow a process to report vulnerabilities: https://simplex.chat/security/
Doing it in public forum hurts your credibility.

2024-05-17 18:22:19

- reply

To reference Signal algorithm specs: https://signal.org/docs/specifications/doubleratchet/#integration-with-x3dh
https://signal.org/docs/specifications/doubleratchet/
It does recommend X448/X25519 algorithms and AES-GCM.
All seems rather boring, isn't it... What's daring here?

2024-05-17 18:18:25

- reply

Ed448/Ed25519 curve choice is controlled by the client, and currently Ed448 curve is not used. What are you specific objections to Ed448 curve?

2024-05-17 18:16:00

- reply

sorry, where did you get 3 days from? It was actually 10 days by 2 people with 5 of them billable (as the report says).
Additional, longer assessment is coming this year.

2024-05-16 21:01:25

Improving communication #privacy requires making many hard choices:
https://simplex.chat/blog/20240516-simplex-redefining-privacy-hard-choices.html#

2024-05-09 21:40:50

- reply

it certainly can be hashed - when user provides the main email and the recovery email, you can validate it. Obviously, user has to remember what the recovery email was for that to work. But a reasonable compromise would be to store hash, salt and x***@***.com pattern so a user could remember which recovery email was used (with or without correct letter count).

2024-05-09 21:38:19

- reply

you do salt hashes at least, right?

2024-05-07 00:10:32

- reply

we have the GitHub issue, but it would be really helpful to have the logs from the device when it happens. Also, having a full list of settings might help reproducing.

2024-05-06 18:15:59

- reply

> E2EE does no good if app is exploited.
I think this is a very wide and generally misleading statement. It's not that black and white. You are talking about different attack vectors, and the purpose of E2EE is to protect from the operator, and the attacks you describe are from untrusted/malicious contacts. Also, the attack via media that would lead to the compromise of E2EE with other contacts is very hard - it's on you to demonstrate it, before stating it's possible.

2024-05-03 10:37:09

We are upgrading the preset SimpleX relays to the new version - it is compatible only with the apps starting from v5.5.3 (released early February) - please upgrade to the latest version and ask your friends to upgrade too.