As far as I know:
Pharo has event recorder that helps automate within image��� however the bigger problem here is interaction with the environment, which is a big problem cross-platform.
This is all assuming that you _must_ click on it and you can't just retrieve-html/parse/extract/download/repeat which is doable in Pharo with Zinc and some non-validating parser.
I believe Kilon worked on some python bindings from Pharo? It would be interesting to ask him what was that exactly��� maybe it could be connected to this lib.
Peter