Perhaps you already have a solution by now but, just in-case here's what we have done in the past.
....We tend to have a steel part integrated into the carbon fiber inorder to produce steel on steel contact for your o-rings. The steel cap can be wound up during the carbon fiber process (assuming you are winding here and not manually laying up). That being said, we leave a little extra material (.060) for a clean up pass on the lathe to ensure a nice sealing surface and to meet tight tolerances requiremnets typically seen in o-ring applications.
I am not sure you will be able to ensure an air tight seal if you are sliding the cap over the tube. But, lets assume you will be able to get the proper distribution of material and a controlled caulk application and assembly process to accomplish this.
You will need to consider both the cohesive and adhesive properties of the caulk you intend to use. The bond strength to both substrates is a value that can be tested and general values should be readily available from the manufacturer. The cohesive strength is a little trickier, but you should also be able to derive the information from the technical data such as tensile strength.
There are a wide variety of materials that you can use, both 1 and 2 part. If air flow is restricted to the caulking, it is probably best to use a 2 part to eliminate the need for air flow. Many caulks require the presence of atmospheric moisture or must flash off a solvent. If you intend to bond the cap to the tube, as you would likely need to to accomplish this, you might be better served by an epoxy or methacrylate compound.