I think you're conflating direct communication with abstract representation. We know that (AFAIK) all primates, and many other animals signal each other, but have failed to find any sign of them creating abstract representations or being capable of learning them despite many concerted efforts to teach them, particularly with primates but other animals too. I think it's clear that these are distinct skills.

Regarding your speculation that it has to involve members of the same species, I don't see any evidence based reason to suspect that and plenty not to. It's looks a lot like special pleading. We've had great success teaching animals to do all sorts of complex tasks. You yourself cite an example of sophisticated inter-species communications and co-operation that argues pretty strongly against the expectation for any requirement for intra-special specificity to communications or learning in general. So given that there don't seem to be any other tasks, skills or communications facilities that require same-species specificity, I'd need to see an argument as to why this one skill is special and might not be communicable across species, when we've not come across such a barrier before.