I don't think there is such a kit that has the capability of separating proteins in a mixture based solely on whether they come from bacteria or insects. There would have to be a property that all the proteins from one source have in common. For example, suppose all the insect proteins were glycosylated in a similar way, but none of the bacterial proteins were glycosylated (completely hypothetical). You might then be able to use immobilized lectin affinity chromatography to separate the insect from the bacterial proteins because the former would bind to the lectin resin and the latter would not. Lacking such a method of differentiation, you would have to use various types of chromatography to purify individual proteins.

There aren't kits to do this.  Depending upon the scale, there may be polyclonal antibodies against total insect digestive proteins or the bacterial proteins that can be used subtractively to clear one kingdom's protein from the other but you would have to hope for non cross-reactivity and expect low recovery which may be fine for analytical applications.  

I am not very sure if this will work but try

1- Rrun the mixture on a gel to separate them based on molecular weight,

2- Do a computer analysis of the gel to get the differents components

3- you may sequence the bands and submit to GenBank for more details

4- Find out which component is a marker for Insects or Bateria

and c'est ca you have your data

Good luck