Thanks for the response, I can use the node you mentioned - and it will work, using the http node and directly querying GitHub and transmitting the file will also work.
However, I would like to know why using filesystem causes errors, but using the memory storage system does not.
So, filesystem will now be default and memory storage will be deprecated. (I believe)
Filesystem mode will only output a relative path if trying to access the binary object. Memory storage will provide the file data. It’s inefficient at large scales.
This will become default and main way to access file objects: