AIKIDO-2026-108299

TL;DR

The rubyzip library checks extracted paths against a destination directory using a prefix match that does not enforce a directory boundary. A crafted ZIP archive can pass this check when the resolved output path shares a prefix with the destination directory, allowing files to be written outside the intended extraction folder. The same release also validates declared central-directory entry counts against header size and stops iterating on the first invalid entry to reduce denial-of-service impact from malformed archive metadata. The fix adds a directory-separator boundary to the containment check and hardens central-directory parsing.