I have no experience whether the network actualy is used during an image build as well, as I always seperate build and runtime configurations. I usualy have a makefile that encapsulates the details on how to build my apps, my containers, publishes their artifacts and deploys thir services. I use the makefile localy during development and from within a cicd-pipelines that is triggered by git pushs.
I only experienced those type of network issues on AWS with EKS and KOPS clusters, when I deploy a Kubernetes build pod to build base images that download stuff from the internet. I never experienced the same problem with Docker Enterprise on AWS.
Thus, said: I have never seen that a network inspect does not list all containers connected to it. it is odd that your container only is attached to a network called “bridge” to it, even though you placed it in a different network. It is absolutly odd that all details have empty/null values for this network. I have never seen this before.
After gathering the facts (main fact beeing: details are missing and wrong), I am clueless as you. It doesn’t make sense and looks like something is seriously messed up. I would disable the firewall, restart the host and see if the behavior still exists - maybee it is just a hickup.
Good luck!