A critical flaw in Red Hat Quay, the enterprise container registry, can let an attacker take over the signing of registry tokens and open up private container images. The vulnerability, tracked as CVE-2026-123710, is scored 9.8 out of 10, with no fixed version published yet.

Quay stores and distributes the container images teams use to build and run applications. The flaw sits in the service that manages signing keys. Where that service is reachable by untrusted clients, an attacker needs no account: with anonymous access switched on, they can obtain a registry token and abuse key rotation to install their own trusted signing key.

Private images and secrets at risk

Once the attacker's key is trusted, they can issue tokens granting read and write access to any repository. That can expose private images and any credentials or secrets inside them, and let the attacker overwrite images other systems later download and run. The malicious key can be set with no expiry date and survives a registry restart unless it is removed.

Red Hat says key rotation deletes the legitimate instance key as a side effect, and the attacker's key can delete replacement keys, disrupting uploads and downloads. Switching off anonymous access does not close the hole, because any valid Quay account can still exploit it, and changing passwords or regenerating robot tokens does not invalidate the attacker's key.

What to do until a fix lands

Red Hat advises restricting access to the key-management service, checking logs for suspicious key changes and removing unauthorised keys, while warning attackers could reinstall malicious keys until a fix is deployed. Fixes for this flaw and a second Quay issue, CVE-2026-110697, were merged into the project's code repository on 11 November; the flaw was discovered by Red Hat's Dave O'Connor.