The container should restrict acquiring additional privileges via suid or sgid bits
Set up the docker integration.
Description
You should restrict the container from acquiring additional privileges via SUID or SGID bits.
Rationale
A process can set the no_new_priv bit in the kernel and this persists across forks, clones and execve. The no_new_priv bit ensures that the process and its child processes do not gain any additional privileges via SUID or SGID bits. This reduces the danger associated with many operations because the possibility of subverting privileged binaries is lessened.
Audit
Run this command: docker ps --quiet --all | xargs docker inspect --format '{{ .Id }}: SecurityOpt={{ .HostConfig.SecurityOpt }}'
This command returns all of the security options currently configured for containers. The option no-new-privileges should be one of them.
Start your container with the options docker run --rm -it --security-opt=no-new-privileges ubuntu bash
Impact
The no_new_priv option prevents LSMs like SELinux from allowing processes to acquire new privileges.
Default value
By default, new privileges are not restricted.
References
- https://github.com/projectatomic/atomic-site/issues/269
- https://github.com/docker/docker/pull/20727
- https://www.kernel.org/doc/Documentation/prctl/no_new_privs.txt
- https://lwn.net/Articles/475678/
- https://lwn.net/Articles/475362/
CIS controls
Version 6
5 Controlled Use of Administration Privileges Controlled Use of Administration Privileges
