That's not evidence of another agent. I can get fog with p-aminophenol. As a rule, a developing agent will be a potentially harmful ingredient that should be reported in the MSDS. Is there any other evidence for the presence of more than one agent? In any case, a developer containing p-aminophenol as the sole developing agent can produce the same characteristics as the current AGFA product, so one might wonder why AGFA would bother to complicate matters.
I agree that the presence of the restrainer isn't conclusive proof of another developing agent. I imagine you probably could induce fog with Rodinal by increasing it's activity. In fact, I'd be interested to hear if the pH of Agfa Rodinal and traditional Rodinal differ substantially.
The reason for the use of a second developing agent, is probably cost-related. p-Aminophenol HCL is considerably more expensive than Hydroquinone and even a bit more expensive than Metol. Exploiting superadditivity with a second developing agent might improve the economics of manufacture.
That being said, 1L of the traditional formula requires about 600 g of three different solutes. Potassium metabisulfite isn't particulary cheap, either, and traditional Rodinal requires 300g in a liter of stock solution.
As you said, why would AGFA bother? All I can come up with is cost to answer that question.
R09 is very dark brown-red, while Agfa rodinal is much more translucent and when diluted pink. It is advantage when used (in small amounts) for positive development. I heard from some people they add sometimes small amount of rodinal to their standard positive developer ...
