What do you mean by "secure"? What's your threat model?

e.g. I don't see how your idea would protect against XSS, or against a heartbleed type bug. What are you trying to protect against?