Red Hat Quay flaw could expose private container images

A critical vulnerability in Red Hat Quay could allow attackers to access private container images, replace repository content and maintain persistent access across an affected registry. The flaw, tracked as CVE-2026-123710, has a CVSS score of 9.8 and affects deployments where the service that manages signing keys is reachable by untrusted clients.

Red Hat’s advisory states that an unauthenticated attacker could seize registry token signing when anonymous access is enabled, by exploiting weaknesses in the key rotation mechanism to install their own trusted signing key. Once in place, the attacker can generate tokens with read and write privileges for any repository, bypassing normal access controls and potentially exposing private images and credentials, as well as overwriting images that downstream systems may fetch.

The malicious signing key can be configured without an expiry and may survive registry restarts unless removed. Red Hat notes that rotation can delete the legitimate instance key as a side effect, while the attacker’s key can delete replacement keys, disrupting normal registry operations. Even disabling anonymous access does not fully mitigate the risk, since any valid Quay account could still exploit the flaw. Changing passwords or regenerating robot tokens does not invalidate an unauthorized signing key.

Red Hat advises restricting access to the key-management service, reviewing logs for suspicious key changes and removing unauthorized keys; however, attackers could reinstall malicious keys until a fix is deployed. The vulnerability was discovered by Red Hat’s Dave O’Connor. While no fixed version is named in the advisory or Bugzilla, fixes for this issue and CVE-2026-110697 were merged into the Quay project on 11 November.