Plain Text Nostr

<-- back to main feed

thread · root f2ee7330…c1ae · depth 2 · · selected dac87ae3…1452

thread

root f2ee7330…c1ae · depth 2 · · selected dac87ae3…1452

50d94fc2d858 -- 172d [root] 
|    story time. the recently disclosed nutshell cashu mint vulnerability is as ironic as it gets. it’s very similar
|    to an inscription which is hilarious. as per the cashu spec, a HTLC must have a preimage witness size of 32
|    bytes.
|    
|    unfortunately, the mint never checked the size before validating and storing it in its db. we simply overlooked
|    it. since users never paid a fee that depends on the witness size (because we assumed it would be constant),
|    this allowed the attacker to store jpgs of dickbutts in a mints database. for free!
|    
|    fortunately there’s no messy consensus in cashu. every mint operator dictates their own rules. the fix is
|    simple: now we reject all tokens with a witness that’s too large. those maliciously crafted tokens (of which we
|    haven’t seen any in the wild) can’t be spent anymore.
|    
|    i must admit, given my recent active engagement in the filter debate, this is probably the funniest exploit
|    possible. i own this one and i’m giggling as i type this. it’s pure comedy.
|    
|    however, this doesn’t mean the disclosure has gone well. the attacker has proven to be malicious and refused to
|    coordinate with us. instead, he’s putting active mints at risk. this is not how responsible disclosure works.
|    very unprofessional. if you run a mint or know someone who does, update to the latest version (0.18.1) where
|    this issue is fixed. funds were never in danger.
|    
|    it’s certainly worth a laugh. grill me. this one is simply too good. 😊
|    
|    thanks to the entire cashu team for their amazing work and their swift reaction. you’ve handled it like pros.
|    reply [1 reply]

Write a post

Sign in with a signing-capable method to publish.