Docker Desktop before 4.5.1 on Windows allows attackers to move arbitrary files. NOTE: this issue exists because of an incomplete fix for CVE-2022-23774.

In Docker before versions 9.03.15, 20.10.3 there is a vulnerability in which pulling an intentionally malformed Docker image manifest crashes the dockerd daemon. Versions 20.10.3 and 19.03.15 contain patches that prevent the daemon from crashing.

In Docker before versions 9.03.15, 20.10.3 there is a vulnerability involving the --userns-remap option in which access to remapped root allows privilege escalation to real root. When using "--userns-remap", if the root user in the remapped namespace has access to the host filesystem they can modify files under "/var/lib/docker/<remapping>" that cause writing files with extended privileges. Versions 20.10.3 and 19.03.15 contain patches that prevent privilege escalation from remapped user.

Docker Desktop Community before 2.5.0.0 on macOS mishandles certificate checking, leading to local privilege escalation.

util/binfmt_misc/check.go in Builder in Docker Engine before 19.03.9 calls os.OpenFile with a potentially unsafe qemu-check temporary pathname, constructed with an empty first argument in an ioutil.TempDir call.

runc through 1.0.0-rc8, as used in Docker through 19.03.2-ce and other products, allows AppArmor restriction bypass because libcontainer/rootfs_linux.go incorrectly checks mount targets, and thus a malicious Docker image can mount over a /proc directory.

Docker Desktop Community Edition before 2.1.0.1 allows local users to gain privileges by placing a Trojan horse docker-credential-wincred.exe file in %PROGRAMDATA%\DockerDesktop\version-bin\ as a low-privilege user, and then waiting for an admin or service user to authenticate with Docker, restart Docker, or run 'docker login' to force the command.

In Docker CE and EE before 18.09.8 (as well as Docker EE before 17.06.2-ee-23 and 18.x before 18.03.1-ee-10), Docker Engine in debug mode may sometimes add secrets to the debug log. This applies to a scenario where docker stack deploy is run to redeploy a stack that includes (non external) secrets. It potentially applies to other API users of the stack API if they resend the secret.

runc through 1.0-rc6, as used in Docker before 18.09.2 and other products, allows attackers to overwrite the host runc binary (and consequently obtain host root access) by leveraging the ability to execute a command as root within one of these types of containers: (1) a new container with an attacker-controlled image, or (2) an existing container, to which the attacker previously had write access, that can be attached with docker exec. This occurs because of file-descriptor mishandling, related to /proc/self/exe.

RunC allowed additional container processes via 'runc exec' to be ptraced by the pid 1 of the container. This allows the main processes of the container, if running as root, to gain access to file-descriptors of these new processes during the initialization and can lead to container escapes or modification of runC state before the process is fully placed inside the container.

The SwarmKit toolkit 1.12.0 for Docker allows remote authenticated users to cause a denial of service (prevention of cluster joins) via a long sequence of join and quit actions. NOTE: the vendor disputes this issue, stating that this sequence is not "removing the state that is left by old nodes. At some point the manager obviously stops being able to accept new nodes, since it runs out of memory. Given that both for Docker swarm and for Docker Swarmkit nodes are *required* to provide a secret token (it's actually the only mode of operation), this means that no adversary can simply join nodes and exhaust manager resources. We can't do anything about a manager running out of memory and not being able to add new legitimate nodes to the system. This is merely a resource provisioning issue, and definitely not a CVE worthy vulnerability.