One possibility is to create a base backlog project that is a Sprint project, but without the time constraints. I think that is what Wikidata plans to do. The Wikidata project will have the special character in the name, but without the time constraints. This project then serves as a container for all project related tasks. Then, as new sub-project Sprints are created, the tasks that have already been triaged and assigned points from the base project can just be added to the Sprint.

What we came up with is writing the number of points in the title of the task itself. Check out the right column "Ready for Dev": https://phabricator.wikimedia.org/project/board/840/?order=priority

At our sprint planning meeting we take tasks from that column and add them to a sprint. After that, it's pretty trivial to go through each task and add the point values to the "Story Points" field.

It would be nice if we could type in the points at our tasking meeting (before we actually meet again to add them into a sprint project). I think having all these tasks in a special sprint could solve the issue.

Thanks @Christopher and @kevinator. One of these approaches should work for us. Appreciate your quick feedback!