OpenPGP Certificate Flooding
My public cryptographic identity has been spammed to the point where it is unusable in standard workflows. This blogpost talks about what happened, what I'm doing about it, and what it means for the broader ecosystem.
If you work with me and you use OpenPGP certificates to do so, the crucial things you should know are:
Do not refresh my OpenPGP certificate from the SKS keyserver network.
Use a constrained keyserver like keys.openpgp.org if you want to check my certificate for updates like revocation, expiration, or subkey rollover.
If you have already fetched my certificate in the last week, and it is bloated, or your GnuPG instance is horribly slow as a result, you probably want to delete it and then recover it via one of the channels described above.
Some time in the last few weeks, my OpenPGP certificate, 0xC4BC2DDB38CCE96485EBE9C2F20691179038E5C6 was flooded with bogus certifications which were uploaded to the SKS keyserver network.
SKS is known to be vulnerable to this kind of Certificate Flooding, and is difficult to address due to the synchronization mechanism of the SKS pool. (SKS's synchronization assumes that all keyservers have the same set of filters). You can see discussion about this problem from a year ago along with earlier proposals for how to mitigate it. But none of those proposals have quite come to fruition, and people are still reliant on the SKS network.
Previous Instances of Certificate Flooding
We've seen various forms of certificate flooding before, including spam on Werner Koch's key over a year ago, and abuse tools made available years ago under the name "trollwot". There's a keyserver-backed filesystem proposed as a proof of concept to point out the abuse.
There was even a discussion a few months ago about how the SKS keyserver network is dying.
So none of this is a novel or surprising problem. However, the scale of spam attached to certificates recently appears to be unprecedented. It's not just mine: Robert J, Hansen's certificate has also been spammed into oblivion as well. The older certification spam on Werner's certificate, for example is "only" about 5K certifications (a total of < 1MiB), whereas the certification spam attached to mine is more like 55K certifications for a total of 17MiB, and rjh's is more than double that.
What Problems does Certificate Flooding Cause?
The fact that my certificate is flooded quite this badly provides an opportunity to see what breaks. I've been filing bug reports and profiling problems over the last day.
GnuPG can't even import my certificate from the keyservers any more in the common case. This also has implications for ensuring that revocations are discovered, or new subkeys rotated, as described in that ticket.
In the situations where it's possible to have imported the large certificate, gpg exhibits severe performance problems for even basic operations over the keyring.
This causes Enigmail to become unusable if it encounters a flooded certificate.
It also causes problems for monkeysphere-authentication if it encounters a flooded certificate.
If this spammed certificate is in the GnuPG keyring, just verifying an
OpenPGP-signed tag in the git revision control
system made by this certificate is now extremely expensive.
-v $tagname, for a tag that is signed with the signing-capable subkey
of this certificate consumes 145 seconds of CPU time (tag signature
verification often happens as part of an automated process, and
typically takes much less than 1 second of CPU time).
There are probably more! If you find other problems for tools that deal with these sort of flooded certs, please report bugs appropriately.
Dealing with Certificate Flooding
What can we do about this? Months ago, i wrote a draft about abuse-resistant keystores that outlined these problems and what we need from a keyserver.
Use Abuse-Resistant Keystores to Refresh Certificates
If the purpose of refreshing your certificate is to find key material updates and revocations, we need to use an abuse-resistant keyserver or network of keyservers for that.
Fortunately, keys.openpgp.org is just such
a service, and it was recently launched. It seems to work! It can
distribute revocations and subkey rollovers automatically, even if you
don't have a user ID for the certificate. You can do this by putting
the following line in
and ensure that there is no
keyserver entry at all in
This keyserver doesn't distribute third-party certifications at all,
though. And if the owner of the e-mail address hasn't confirmed with
the operators of
keys.openpgp.org that they want that keyserver to
distribute their certificate, it won't even distribute the
certificate's user IDs.
This keyserver also doesn't have the same keys as the SKS pool. It was seeded with the keys on the pool on setup, but is not pulling new updates in nor sending updates back.
Fix GnuPG to Import certificate updates even without User IDs
Unfortunately, GnuPG doesn't cope well with importing minimalist certificates. I've applied patches for this in debian experimental (and they're documented in debian as #930665), but those fixes are not yet adopted upstream, or widely deployed elsewhere.
In-band Certificate Discovery
Refreshing certificates is only part of the role that keyserver networks play. Another is just finding OpenPGP certificates in the first place.
The best way to find a certificate is if someone just gives it to you in the context that it makes sense.
The Autocrypt project is an example of this pattern for e-mail messages. If you can adopt an Autocrypt-capable e-mail client, you should, since that will avoid needing to search for keys at all when dealing with e-mail. Unfortunately, those implementations are also not widely available yet.
Certificate Lookup via WKD or DANE
If you're looking up an OpenPGP certificate by e-mail address, you
should try looking it up via some mechanism where the address owner
(or at least the domain owner) can publish the record. The current
best examples of this are
and DANE's OPENPGPKEY DNS
records. Modern versions of
GnuPG support both of these methods. See the
This is a mess, and it's a mess a long time coming. The parts of the OpenPGP ecosystem that rely on the naive assumptions of the SKS keyserver can no longer be relied on, because people are deliberately abusing those keyservers. We need significantly more defensive programming, and a better set of protocols for thinking about how and when to retrieve OpenPGP certificates.
A Personal Postscript
I've spent a significant amount of time over the years trying to push the ecosystem into a more responsible posture with respect to OpenPGP certificates, and have clearly not been as successful at it or as fast as I wanted to be. Complex ecosystems can take time to move.
To have my own certificate directly spammed in this way felt surprisingly personal, as though someone was trying to attack or punish me, specifically. I can't know whether that's actually the case, of course, nor do I really want to. And the fact that Robert J. Hansen's certificate was also spammed makes me feel a little less like a singular or unique target, but I also don't feel particularly proud of feeling relieved that someone else is also being "punished" in addition to me.
But this report wouldn't be complete if I didn't mention that I've felt disheartened and demotivated by this situation. I'm a stubborn person, and I'm trying to make the best of the situation by being constructive about at least documenting the places that are most severely broken by this. But I've also found myself tempted to walk away from this ecosystem entirely because of this incident. I don't want to be too dramatic about this, but whoever did this basically experimented on me (and Robert) directly, and it's a pretty shitty thing to do.
If you're reading this, and you set this off, and you selected me specifically because of my role in the OpenPGP ecosystem, or because I wrote the abuse-resistant-keystore draft, or because I'm part of the Autocrypt project, then you should know that I care about making this stuff work for people. If you'd reached out to me to describe what you were planning to do, we could have done all of the above bug reporting and triage using demonstration certificates, and worked on it together. I would have happily helped. I still might! But because of the way this was done, I'm not feeling particularly happy right now. I hope that someone is, somewhere.