Podman vulnerabilities

Podman from 4.4.0 until 5.8.8 and 6.1.3 treats an image carrying the io.podman.annotations.checkpoint.runtime.name annotation as a checkpoint and silently ignores user-supplied container creation restrictions such as --cap-drop=ALL. This allows image-controlled checkpoint state to select weaker sandbox settings and capabilities. Rootless containers are not affected because checkpoint and restore are unavailable in that mode. This issue is fixed in versions 5.8.8 and 6.1.3.

A flaw was found in Podman. If an attacker can pass a crafted tar archive to the `podman load` command, they can create files on the host machine with the privileges of the user running Podman.

The 'podman quadlet install --replace' command opens the existing Quadlet file without O_TRUNC, which on filesystems without reflink support (e.g., default XFS on RHEL) leads to incomplete overwriting. If the new Quadlet is shorter than the original, old data remains at the end of the file, and the command completes without warning.

Podman from version 1.8.1 to 5.8.4 contains a vulnerability where a container image with an environment variable consisting only of a key (no value) can trick Podman into passing that variable from the host into the container. Using an asterisk (*) as the key causes all host environment variables to be passed, allowing a malicious image to exfiltrate Podman environment variables from the session where the container is launched.

Podman versions 4.8.0 through 5.8.1 contain a command injection vulnerability in the HyperV backend, where the VM image path is inserted into a PowerShell string without sanitization, allowing arbitrary PowerShell commands to be executed with Podman process privileges (SYSTEM on Windows).

A flaw in Podman causes data written to `RUN --mount=type=bind` mounts during `podman build` to persist, making files created inside the container appear in the host's temporary build context directory.

A vulnerability in Podman allows an attacker to overwrite host files using the 'kube play' command when the kube file contains a Secret or ConfigMap volume mount with a symbolic link to a host file path. The attacker can control which file is overwritten but not the content written.

A flaw was found in Podman where the `podman machine init` command fails to verify the TLS certificate when downloading VM images from an OCI registry. This enables a Man-in-the-Middle (MITM) attack.

A vulnerability in `podman build` and `buildah` allows container breakout via a race condition with --jobs=2 when building a malicious Containerfile. Even with SELinux enabled, it allows enumeration of files and directories on the host.

A flaw in Podman allows an attacker to create a specially crafted container that, when sharing IPC with another container, creates a large number of IPC resources in /dev/shm. The malicious container exhausts resources until OOM-killed, but the IPC resources remain in the IPC namespace, and with automatic restarts this leads to memory exhaustion of the system.