As my list of actions grows, I'm finding that the "checkbox" interface for applying actions isn't always very convenient. Since the list is long, many actions scroll off the bottom and you always have to be careful to see whether any are checked, otherwise you get unexpected (or worse) results.
I'm trying to think of what a better approach would be... One thing I'd like is the ability for a user defined action to include other user defined actions. This would give the ability to create a complex action and use it within more than one other.
So instead of doing something like:
[x] Action A
[x] Action B
[ ] Action C
[x] Action D
[ ] Action E
[x] Action F
You might have Action A call Action B, D and F. Similarly, Action C could call those same actions. This would be analogous to using subroutines in a procedural language.
Now, with this ability, the "sub"actions might not need to appear in the interface. So also give the user the ability to 'hide' any unchecked actions. In the example above, the user might hide B, D, and F so that the action list in the interface isn't as long.
(By the way... While we're on the subject, do you think it would be wise to rename the "actions" that are created by users, so that they're differentiated from the atomic "actions" from which they are created? Maybe call the user defined actions "scripts" or something similar.)