search

w3c / vc-data-model

W3C Verifiable Credentials v2.0 Specification
https://w3c.github.io/vc-data-model/
Other
299 stars 106 forks source link

Clarification on the Role of Verifiers in Preventing Replay Attacks #1548

Closed laysakura closed 3 months ago

laysakura commented 3 months ago

https://github.com/w3c/vc-data-model/blob/4c6005da8e44ee0d78178b567989481b6012ee6e/index.html#L6038-L6041

The current wording in the document is technically accurate but may be misleading.

In the scenario described, the attacker is the holder. It seems unlikely that a malicious holder would voluntarily include a nonce or expiry timestamp in a Verifiable Presentation (VP) to prevent replay attacks.

It would be more appropriate to phrase this section with the verifier as the subject, suggesting that the verifier may require a nonce or expiry timestamp to mitigate the risk of such attacks. This adjustment would clarify the roles and responsibilities in this context and make the document more accurate.

msporny commented 3 months ago

Yes, that section is strange and needs to be updated. I expect that it said something different long ago and and based on a series of changes over the years, does not make much sense today.

A PR should be raised to do exactly as you say -- the verifier should ask the holder to include a challenge, nonce, or expiry timestamp to mitigate the risk of replay attacks. We should also specify when a nonce is useful (where the holder wants to include extra information to ensure their signature isn't replayed, if the verifier enforces nonce uniqueness (which it should).

laysakura commented 3 months ago

@msporny Got it. I created a PR for this: #1550

msporny commented 3 months ago

PR #1550 has been raised to address this issue (Thank you, @laysakura!). This issue will be closed once PR #1550 has been merged.