Repository navigation
Replies: 5 comments 12 replies
|
Okay, I only have some official pico2 as spare keys. Are they sufficient for the testing requirements? |
|
It is a good idea, but the importance here is not to over-advertise this as a security improvements. It shrinks the attack surface a bit, but it is still not a secure element as such. P.S. out of context, but related: have you researching attaching a separate SE element via I2C? |
|
Has anyone found any other issues in the new firmware? |
|
Thanks for the ping. Wish I could help, but I've only got one RS-Key board (the Waveshare RP2350-Touch-LCD-2.8 V2, CST3530 — I added the V2 support earlier), and it's my daily driver with production creds on it, so like you said, I'd better not flash develop on it. I'm also short on time the next couple of weeks. I'll update to 0.5.0 once it's released and check the V2 display/touch path — I'll report back here if anything's off. Also, since you mentioned sourcing parts from China — I'm based there, so if you ever need help with suppliers, buying or forwarding stuff, just let me know. Good luck with 0.5.0. |
|
A small update from us: we have no comments on the main functionality, and basic pentesting didn’t turn up anything either. Since @Babel36acl mentioned the model with the LCD screen: I’d propose an experimental on-device transaction confirmation feature. The key would show the payment details (amount, payee) on its own display and include them in what it signs, i.e. “what you see is what you sign”. thirdPartyPayment from CTAP 2.2/2.3 is already in develop, but it only marks the credential, and the details are shown by the browser (SPC). The display part is still only a WebAuthn proposal (w3c/webauthn#2020), so this would be a hands-on prototype of where the spec might go. It’s purely academic interest, as I doubt banks would ever rely on RP2350-based security keys. |
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
@token2 @lockedmutex @cresseelia @netll @mannp @Curious-r @ncmro7 @lpiob @JDestiny @Babel36acl
Hi everyone. For the past two months, I’ve been working on a major update: the firmware now supports a Secure/Non-Secure split via TrustZone, which is enabled by default. Before I push a new release, I’d like you to test this version if you have the capacity to do so. It’s available in the
developbranch, and both the documentation and the build process have been updated accordingly. I recommend testing it first on boards that do not contain any of your production secrets. Ideally, you should test all the applets you use, but please do so using test or backup data—data that you can afford to lose or otherwise recover.I anticipate the testing phase will last for two weeks (until November 21, 2026). Afterward, I will address any remaining issues or bugs and prepare the final 0.5.0 release. If you encounter any problems, feel free to post in this thread or open a separate issue—whichever you prefer; either way is fine. Thanks for your time and attention.
Target branch: https://github.com/TheMaxMur/RS-Key/tree/develop
Docs:
https://themaxmur.github.io/RS-Key/develop/build.html
https://themaxmur.github.io/RS-Key/develop/architecture.html
https://themaxmur.github.io/RS-Key/develop/production.html
All reactions