Some news for you:
* AtlasStyler has been developed April to July 2008.. and since then didn’t get a lot of new features.
* My main focus on development in the last months was on Geopublisher, less on AtlasStyler.
* I switched from GeoTools 2.4 to latest Geotools 2.6. Soon so be released AtlasStyler version 1.3 will profit from that.
* A lot of internal reorganization and cleanup was made. If you look at http://www.ohloh.net/p/Geopublishing/analyses/latest, you can see that the lines of code are stagnating, even though many feature commits were made. I see this phase as very important to continue with a clean codebase.
* I do not test any AtlasStyler SLD with Geoserver myself – as a developer i have few time for proper testing. And i don’t use Geoserver.

I read about your problems with AtlasStyler and Geoserver. Please report them! AtlasStyler and Geopublisher could be a great pair.. they use the same internals. It just needs some power-users that find some bugs and smooth they could go.

Please consider postiing your problems to the users mailing list or add them to the bugtracker.
AtlasStyler user list: http://lists.wald.intevation.org/mailman/listinfo/atlas-framework-users
AtlasStyler Bugtracker: http://wald.intevation.org/tracker/?group_id=37

About PostGIS connectivity… i expect it not to be hard, as Geotools supports it very well. I would maybe take a week of programming until it’s stable. WFS is a bit trickier.. maybe 1 to 2 weeks of work.

I invest full-time into Geopublisher and AtlasStyler, so please understand that i can not write all feature requests unpaid. If your organization profits from the use of AtlasStyler, please consider investing into this free open-source project. Make an offer

A good news at last:
One of the main new AtlasStyler features in version 1.3 are better labelling options, as described in a blog entry.
http://de.geopublishing.org/AtlasStyler%20SLD%20Labeling%20Tutorial