There wouldn’t be any need to wait there, as you simply adjust the amount of time spent at Hirwaun accordingly. So long as the running time on the extension is no more than 15 mins each way it would work nicely.
Yes, the two trains would need to pass at Aberdare, but I seem to recall that the infrastructure is there already, albeit not a second platform, immediately north of the current station.
According to my Quail map (which is some years old so things may have changed) there is a siding just north of Aberdare but no loop. The route looks unobstructed so this could probably be made into a loop and extended if necessary so the trains don't have to stop when passing each other. Aberdare platform looks to be blocking the former second track so a second platform there would be difficult.
There is a non-platform loop at Abercwmboi a mile and a bit to the south with an intervening station at Cwmbach. If the timetable is such that trains pass at Abercwmboi then they will probably need about 5min from there to Aberdare (including the Cwmbach stop), 5min at Aberdare and 5min back again. In this situation converting the siding just north of Aberdare into a loop looks close to the optimum, because it would be about 5min from a northbound train arriving at Aberdare, for it to go into the loop and the corresponding southbound one to move from the loop to Aberdare, make its call there and leave on time. To avoid the need for another loop the train would have to get from this loop up to Hirwaun, wait there, and come back (including any new intermediate stops) all within 25min. I suspect if the trains passed at Mountain Ash instead, then this would need two extra loops even for a 30min interval at Hirwaun.
There's also a trainman-operated level crossing at Robertstown which is close to a roundabout and very tight bend so risky to convert to an automatic type. A frequent passenger service would be incompatible with this method of operation and under Network Rail rules the increased risk would almost certainly lead to closure. Tramway operation might help here, especially as this might be a good site for an intermediate station.
So there is a relatively cheap option if the future timetable is as I have assumed, but even that will require significant track and signaling work.