6 ms·

How are you envisioning phishing working here? The credentials are bound to a set of origins. I don't know of any passkey managers that have an option to send a

How are you envisioning phishing working here? The credentials are bound to a set of origins. I don't know of any passkey managers that have an option to send a passkey to the wrong origin because it just doesn't make sense.

namaria 3y ago
I'm sorry but "I don't see it" is not a good argument in security. Hailing a technology as "exploit proof" should be considered a non-starter. Any security evaluation should be about thread models, attack surfaces, mitigation and trade-offs. Never "trust us". Not knowing how a technology could be exploited is not a feature, is a liability.
- nonrandomstring 3y ago
  I think I am seeing a real clash of mind-sets in this thread. People from a security mindset like you and I meet a lot of friction with developers who are chirpy and optimistic (and also more generally with people who want technology to be fluffy and "just work") Not that chirpy optimism is a fault at all, we all need a little more of it, but when it crosses borders into strong beliefs that actually put others at risk I feel discomfort, and a need to challenge those ideas. Now, not understanding complex and subtle security issues is nothing to be ashamed of, and it's wrong (arrogant) if we decry that. What is unacceptable though, is "reckless and dismissive optimism", whether that's in systems, ostensible authorities, technologies or whatever.
  - namaria 3y ago
    Well put. I face a lot of emotional push back when people claim "this is perfectly safe because I don't see how it could be breached". There's probably a component of pride in that people assessed the situation and decided to put trust in what they believe to be sound reasoning. In adversarial dynamics, sound reasoning, as perfect as one can make it, is never enough. Defense takes trade-offs in convenience and imposes costs that people are more then willing to forego by depositing trust in an entity they can hold liable in court later.

namaria 3y ago

I'm sorry but "I don't see it" is not a good argument in security. Hailing a technology as "exploit proof" should be considered a non-starter. Any security evaluation should be about thread models, attack surfaces, mitigation and trade-offs. Never "trust us". Not knowing how a technology could be exploited is not a feature, is a liability.

nonrandomstring 3y ago
I think I am seeing a real clash of mind-sets in this thread. People from a security mindset like you and I meet a lot of friction with developers who are chirpy and optimistic (and also more generally with people who want technology to be fluffy and "just work") Not that chirpy optimism is a fault at all, we all need a little more of it, but when it crosses borders into strong beliefs that actually put others at risk I feel discomfort, and a need to challenge those ideas. Now, not understanding complex and subtle security issues is nothing to be ashamed of, and it's wrong (arrogant) if we decry that. What is unacceptable though, is "reckless and dismissive optimism", whether that's in systems, ostensible authorities, technologies or whatever.
- namaria 3y ago
  Well put. I face a lot of emotional push back when people claim "this is perfectly safe because I don't see how it could be breached". There's probably a component of pride in that people assessed the situation and decided to put trust in what they believe to be sound reasoning. In adversarial dynamics, sound reasoning, as perfect as one can make it, is never enough. Defense takes trade-offs in convenience and imposes costs that people are more then willing to forego by depositing trust in an entity they can hold liable in court later.