A newly disclosed vulnerability, tracked as CVE-2026-17106, enables malicious containers to escape their designated hosts via `docker cp`, overwriting files on the host running Docker CLI This article explores malicious containers escape. . This flaw doesn't grant full host control; it merely converts copy-out operations into a host-side write primitive, executed with the user or automation process's permissions.
Imperva researchers discovered this issue in Docker’s archive pipeline and categorized it as a destination-escape flaw within the `docker container cp` command. Security depends on two key requirements: ensuring the archive accurately reflects the consistent source filesystem and preventing any archive entries from resolving outside the destination.
Docker's Tar Extraction Flow (Source: Imperva) A malicious process in a compromised container can manipulate paths during Docker’s archive creation phase, initially recording them as directories before treating them as symbolic links. A hostile container can target shell startup files, SSH configuration, source trees, executables, cloud settings, or LaunchAgent persistence files that the local user can modify. Until patches are deployed, avoid copying from running or compromised containers to sensitive systems, use sudo-based copy workflows judiciously, enforce least privilege in CI, and collect forensic material through isolated virtual machines or disposable accounts.
Enhance your SOC by gaining complete visibility into phishing activities and reducing Mean Time To Remedy (MTTR) with comprehensive security measures.
