what about the outbox?

Reply to this note

Please Login to reply.

Discussion

Outbox relay connections can spread through interactions and follows. If a client is vulnerable but a "trusted" relay was able to filter/block harmful events, an attacker can convince my client to connect directly to an malicious relay and possibly compromise my client.

is this a common knowledge in nostr relay operators and client devs?

Kind of. It's sort of a priority/trade off thing. Basic XSS can be easily mitigated, but the nature of nostr clients is that they're always fetching unknown and untrusted data from servers (relays) and it's often not a priority of the developer (or their AI vibe sessions) to consider the "what could happen if".

Unfortunately that’s true for many, and don’t get me started about random web extensions, that store your nsec in plain text in browser storage

Which is why I started this https://github.com/VnUgE/NVault. But it's far from being suitable. Still better than storing nsec in the browser though.

I get 404

No worries heres my central site. Most people only click on github links so that's what I share.

https://www.vaughnnugent.com/Resources/Software/Modules/NVault

I am making a highly secure extension addition for the Nostr Build Shack now, should hit public test release this weekend.

https://testflight.apple.com/join/qgkAMPgU Nostr Build Shack

I don't use a smartphone but Ill take your word for it! I'm wanting to shift toward commercial/enterprise customers with on-prem requirements.

Fair game, but it’ll be tough

i stumbled around and eventually landed on and read this post: https://www.vaughnnugent.com/blog/d9ab8a46cfa8d6bd59cf048fec8d73ffc44f881c πŸ‘πŸ‘

Interesting write up

Agreed. Though if I'm being perfectly honest, the approach of "everything is insecure, so I should build my own tools - they'll be more secure!" can be an anti-pattern in the wrong hands. I don't know enough about nostr:npub1qdjn8j4gwgmkj3k5un775nq6q3q7mguv5tvajstmkdsqdja2havq03fqm7 or his work to say, but it's a very minor alarm bell.

Sometimes the whole ecosystem is improved more by talented and motivated individuals contributing to exiting projects that need help in the areas those individuals identify as weak points.

I tried to find clean and compliant libs for swift, just to end up writing my own on top of the standard libsecp256k1 instead of gluing together all the slop

This project doesn't already exist. Noscrypt was built so that others (including myself) are re-rolling their own nip44 crypto every time they want to create a messaging app. In the case of noscrypt the goal is to be as ubiquitous as openssl. Which btw, noscrypt is not hand-rolled, have a look.

Yes there is a judgement cast on "Im smarter and I can build it better" I don't have a good argument to that, it is what it is. My attempt was naivley was to help prevent that. Noscrypt is an "old" project relatively speaking, it is the existing project.

I'll spare my soapbox of a the mono-language/high-level culture language speech, but there is a reason noscrypt is written in a systems language and believe this is the only correct way to be implemented.

I totally agree with you.

This blog is also a long-form note on a relay somewhere :)

I am bookmarking it. would be great to hear an update though when everyone can use it and test it?

Me too XD. Unfortunately it's on the backlog for now, I've got a git server to get up and running XD

It seems you and nostr:nprofile1qqsglv2qkn5dmmuhee9cy8fywfu2rfp4xd3xy0myqg2gfvmjl9yqqrqppemhxue69uhkummn9ekx7mp0qythwumn8ghj7un9d3shjtnwdaehgu3wvfskuep0qythwumn8ghj7un9d3shjtnswf5k6ctv9ehx2ap0crg0k2 are developing similar things. Are you two in collab? ☺️

Similar but not cross compatible (can be in the future)

I am sure there is plenty of room for it when you sched allows it lol 🀣

🀣

Just and http server with an API so just about anything can be (assuming I get alternate/standard session protocol established)