Sooner or later, in every discussion of open science, someone will say there's a risk of being scooped if you share your code or data with your peers. I think this is pure FUD: in all the years I've worked with scientists, I have never met anyone this has actually happened to.
But absence of proof is not proof of absence, so I'd like to issue a challenge. If someone has ever published a result you were going to by taking advantage of software or data that you made publicly available, I'll send you a Software Carpentry t-shirt. You'll need to provide specifics, but I won't share those details with anyone without your permission. Please mail me if you'd like to chat.