Do you still remember PGP 2? When you added a key or changed trust parameters, it would automatically detect that some keys got certified, and would ask you whether you considered these keys' holders trustworthy. GnuPG is lacking this feature: You'll have to use the key editing menu in order to adjust trust levels of any keys for which you consider this appropriate. Of course, this is cumbersome - in particular when you add keys to your key ring routinely, and keeping track of trust levels is not an option.
This script may help you to assign trust levels: It will look for keys on your keyring which carry well-certified user IDs, but haven't had any owner trust level assigned to them. It will then run gpg --edit over all these keys and allow you to nicely adjust trust levels.
As was reported in comp.mail.mutt, there's an interoperability problem between mutt and Eudora when it comes to PGP/MIME encrypted messages. Apparently, Eudora expects that the first MIME header within the encryption envelope contains a MIME-Version: 1.0 field. This is a bug. For details, read the Usenetr thread.
Lance Spitzner of the Honeynet Project writes: The Honeynet Project is excited to announce the release of a new paper, "Know Your Enemy: Statistics". The paper is based on eleven months of data we have collected, and covers two areas.
1. Aggressiveness. We show a variety of statistics, based on eleven months of Honeynet data, that demonstrate just how aggressive the blackhat community can be. Even more revealing is it appears the threat is getting worse.
2. Early Warning and Prediction. We present proof of concept that attacks can be predicted. The research is still preliminary, but based on statistical analysis, 7 out of 8 attacks could be predicted last year. Two different team members took two different statistical approaches, yet they findings were almost identical. Most of the attacks made against the Honeynet could be predicted two to three days in advance.
As this research is part of the Honeynet Project, all eleven months of data are also publicly released for review and further analysis.
An extensive analysis of the Code Red Worm is available here. Includes animations.
When you sign a (public key, user ID) couple using a V3 RSA key, gnupg will create a V3 signature. This includes the case of local signatures. Bad enough, there is no such concept as a local V3 signature, which implies that using "lsign" with a V3 RSA key is equivalent to "sign", and produces an exportable(!) signature.
This, in turn, can lead to considerable embarrassment (and loss of reputation) of the signer, and to bad judgement by others who rely on such signatures.
This patch fixes this problem by adding a minimum version parameter to make_keysig_packet(), and makes sure that local signatures are always at least V4.