If this is happening synchronously in the same process, the child should be able to read the file. So, does it happen synchronously? Is it in the same process?

Does the child (ScriptB) drop privileges? Does it run under a different chroot? How does it try to open the file?

Either the child or PB::FileUtils is doing something here to make the file (seem?) inaccessible, but you haven't shown us that particular code, so I cannot say what.

If you would try running the child with system {@pmcmd[0]} @pmcmd, as I suggested, we might be able to find whether or not the "something" in question is found in PB::FileUtils. But even then, I would have to see the code from the guilty party (be it PB::FileUtils or ScriptB) before I can say what precisely is wrong.