GP Binary Classification Pseudo Code

I am working with various inference methods for classification. Currently I am trying to combine MCMC for producing samples from a posterior distribution with with a Gaussian Process binary classifier. However, I am struggling at how a pseudo code may look. I have successfully implemented the MCMC algorithm and have generated the relevant posterior samples. Now I don't know how to implement the GP classifier once I have these samples. The co-variance function is the squared error function and the mean function is zero.

Any help/tips would be greatly appreciated