I released in 2003 the first few rules to detect unwanted messages (spam) in Portuguese for SpamAssassin. These rules ended up being included in some Linux distributions. I’ve tried to update the project in 2006, but I was unable to keep up the project because of lack of time.
While I was setting up this new personal site, I noticed a lot of hits looking for these old spam rules. So, I decided to keep the original files here so they can be downloaded, even without any updates since 2003.
Setting up these rules means getting a little creative with the set of regular expressions (RegExp) to set a score on the match of each rule. Thus, the sum of these scores classify the message as spam or not (each sysadmin defines its cut-off score).
This project was a great way to learn regular expressions and I recommend it to anyone who wants to learn more about regular expressions. Anyway, I do believe today more advanced machine learning tools should have better performance and accuracy.
I also found some people putting together these rules on other projects like this repository and others.
Download the rules here: br_rules.cf