You can configure Barca's builtin junk mail filters to do this for you or you can build your own using Barca's filtering and a technique for
using files within filters. You could even go further and implement a scripted solution if you liked (this would be considerably more complicated).
I will explain the builtin Junk Mail Filtering (JMF) to see if that works for you. There are two aspects to the builtin JMF filters, Bayesian (see below) and standard non-Bayesian. I will describe the non-Bayesian aspect now as that is what you have asked about.
You first have to enable the JMF. To do this:
- Open the Junk Mail Filtering window by selecting "Tools | Junk Mail Filtering..."
- Check the "Enable automatic Junk Mail filtering" box
- Select a sensitivity level (I recommend High)
- Go to the "General Settings" tab
- Select to run the junk mail filters AFTER downloading messages
- Check the "Run standard non-Bayesian filters" box
- Go to the "Word Lists" tab
- Use the various buttons to enter the criteria you want to use.
Next, open the "Filters" form (Tools | Filters) and select the "Incoming" tab. Make sure the "run Junk Mail filters" entry is at the end of the list (drag it to the end if necessary).
Note, the lists you build here can also be used in standard filters if you want to disable the JMF filters and use your own custom ones.
Bayesian filters are filters that learn over time. They start out fairly dumb and require you to classify messages as good and bad. Barca's implementation requires that you train a certain number of words into what are known as good and junk corpuses. This is done by simply using the "File as Junk" and "Classify as Good" icons at the far right of the preview pane toolbar. People report varying degrees of success with Barca's builtin Bayesian filtering.