Hi @Nils-Nilsson ,
Thanks for the feedback; we discussed this with the team some more.
The use case makes a lot of sense, especially with risk profiles introduced in ProGet 2026. Instead of exposing Policies (which have a bunch of package-driven rules) and blocking, our thoughts are:
[1] Allow risk profile to be set/configured on Docker feeds; behind the scenes this would create Policy, but none of the rules would be used
[2] Allow scoped assessments on Docker feeds; this would also use the feed's policy
[3] Add pgutil containers audit command, which would display vulnerabilities and error if any are severe
The issue we have with "blocking" downloads directly is that the Docker/Kube/Podman/etc. do not expose HTTP error messages and this becomes very painful to debug/troubleshoot.
However, running pgutil containers audit can "break" any automation process (nonzero exit code) and will list all vulenrabilities etc.
LEt us know your thoughts!
Thanks,
Alex