Have experts review our revocation mechanism of Tails signing key
This ticket is about sending https://tails.boum.org/doc/about/openpgp_keys/signing_key_revocation/ to a bunch of smart people and ask them to review and comment on it.
We mentioned dkg.
#10 Updated by intrigeri over 2 years ago
Are you looking for review of the of the cryptography itself, or the threat model?
I doubt the biggest problems of this mechanism lie in the crypto being used, but IMO generally auditors should take developers' intuition with a grain of salt, and look for problems wherever they think they might find any :)
#15 Updated by dkg almost 2 years ago
It would be good to know what kind of review you're looking for. just an e-mailed response that will never be published? some sort of public review, comparing it to other policies? suggestions for improvements in the form of bug reports? plaudits for media consumption?
all of these things are pretty different from each other, so just a generic "asking for review" might be improved with more details.
#16 Updated by sajolida almost 2 years ago
It would be good to know what kind of review you're looking for.
just an e-mailed response that will never be published?
some sort of public review
That's also possible but a less formal review works as well.
Sending a mail to tails-project@ would be in-between a very formal
review and an email that will never be published and work as well (maybe
that would be our preferred option in terms of cost-benefit for the
reviewers and the transparency of the process).
comparing it to other policies?
That would be super interesting though we didn't think of that so far.
Maybe pointers to other similar policies would be good as a start.
suggestions for improvements in the form of bug reports?
That would be more work for the reviewers and I don't think that's needed.
plaudits for media consumption?
Not really :)
The goal here is more to fix issues in the current document while
putting as little overhead on the reviewer's shoulders as possible.
#19 Updated by cypherpunks about 1 year ago
While this has been resolved, I would like to point out that it would be much better if the signing key were stored on a smart card or a HSM. Simply storing it on an encrypted drive makes it much easier to exploit the revocation key-holders. Keeping it on a cheap smart card would not be too much to ask for the key-holders and would greatly improve security.
#20 Updated by intrigeri about 1 year ago
While this has been resolved, I would like to point out that it would be much better if the signing key were stored on a smart card or a HSM.
See https://tails.boum.org/doc/about/openpgp_keys/#signing. The full master key is not stored anywhere and the subkeys are stored on smartcards :)
- Status changed from Resolved to In Progress
- Assignee deleted (
- % Done changed from 0 to 50
(email backlog, yeah.)
I consider this done
Well, not entirely: we've not acted at least on one piece of feedback
<email@example.com> we've received. Paraphrasing: we currently have no way to ensure that folks still have access to their share and that the place where it's stored is still considered safe. So reopening this ticket (asking folks to review is one part of it, ensuring we take their feedback into account is another one). Suggested action: set up a process to email them all every two years.
I mean, this should be a continous process.