Open kristovatlas opened 8 years ago
I've seen more than one wallet vendor (I won't name names since it's not relevant here) make claims similar to "Your transactions are anonymous." without any qualifiers. I'm not 100% sure of how to precisely quantify this as a criterion, but I think it's reasonably accurate to say that there aren't any blockchain systems that are definitely anonymous with no qualifiers, and that any wallet provider that makes such claims is harming their users' privacy (if nothing else because users who are told that they are anonymous with no qualifiers are psychologically more likely to do risky things).
There are more specific risks as well which should be disclosed, but I think the nonexistence of absolute anonymity should definitely be on the list.
attack - mislead users about privacy properties
The attack that has been added in the PR:
Hide adverse privacy behavior from users by not disclosing or by misrepresenting privacy risks.
The obvious countermeasure is:
Disclose privacy risks to users in a public location.
Proposed criteria:
Unable to come to consensus about the countermeasure. Moving milestone.
The attack will be left with an empty list of countermeasures for 3rd edition. We discussed this attack tonight and decided to try to "penalize" wallets who obviously publicly misrepresent their privacy in the blurb describing the wallet in the next report, rather than factoring it into the 3rd edition score.
We'll revisit the issue of countermeasures/criteria for 4th edition of the threat model.
@wtogami suggested that we consider the degree to which wallet providers disclose risks to users transparently. For example, it might be helpful to let users know what party service is used for lookup queries, if one is used.
A sample "attack" might be something like: "The wallet provider misleads users about or neglects to inform users about risk X" with the "countermeasure" being informing users about risk X.
The task is to come up with a list of risk X's that we think are worth disclosing. We could also consider degrees of effectiveness of disclosure, though this is added complexity. A place to start is to simply make it binary.
Relevant criteria would likely have to be gathered in a questionnaire response from wallet providers.