I agree with the assumption that people could be easier attracted to this project if they have some "toys" to play with. Some may have good ideas for parts of a bot but are afraid of other parts, e.g. the navigation. Through this proposed plugin system they could concentrate on their part and don't have to worry about other parts. In the end this should result in a larger development community.
Possible first goals:
- spawn a bot with the same API in more then one game engine (as PMB suggested)
- concentrate on ONE game engine and start with the design of some of the plugin interfaces, e.g. for an aiming plugin.
I should start to derust my C++, havn't used it much in the past years...