Skip to content
← All stories

Vulnerabilities · Cloud Security · Application Security

Critical Red Hat Quay flaws let attackers steal and overwrite private container images

Two critical flaws in Red Hat Quay could let attackers read and overwrite private container images. Red Hat has released fixes, but all Quay 3 and 5 versions are affected.

Ross Grant3 min read

Laptop installing a security update

The short version

Red Hat Quay is a tool that stores container images for software development. Two serious security flaws mean attackers could steal or change private images in any Quay registry. Red Hat has released fixes, but all current versions are affected. If you use Quay, you should update as soon as possible.

By the numbers

  • 9.8CVSS score for CVE-2026-123710The Hacker News
  • 9.9CVSS score for CVE-2026-110697The Hacker News
  • 2Number of critical flaws disclosedThe Hacker News

What happened?

Two critical flaws in Red Hat Quay allow attackers to read and overwrite private container images in any repository. The flaws affect all Quay 3 and Quay 5 versions, and Red Hat has released fixes. Attackers can exploit the main flaw even if anonymous access is disabled, according to The Hacker News. Only one outlet has reported this so far.

Who is affected

Every version of Red Hat Quay 3 and Quay 5 is affected by both critical flaws, according to The Hacker News (opens in a new tab). This includes standalone installations, Quay deployed by the Quay Operator on OpenShift, and the mirror registry for Red Hat OpenShift. Any Quay instance that can be reached by untrusted clients is exposed, including internal registries. Quay enables anonymous access by default, but even with it disabled, anyone with a valid account can exploit the main flaw.

How the flaw works

The first flaw, CVE-2026-123710 (CVSS 9.8), lets an attacker add their own key to the list of keys Quay trusts for signing access tokens. These tokens allow users to push and pull container images. Once the attacker adds a key (with no expiry date), they can create tokens that let them read or overwrite any private image in the registry. This includes images that may contain secrets or credentials. The attacker can also delete the keys used by running Quay instances, which would cause normal authentication to fail for legitimate users. The planted key remains active until it is manually removed, and it survives restarts.

The second flaw, CVE-2026-110697 (CVSS 9.9), allows a user who can create a namespace to steal robot account tokens from other tenants. This is due to a bug in how Quay matches namespaces using SQL patterns. By creating a namespace with an underscore in the right place, an attacker can access robot accounts belonging to other users, along with their decrypted tokens. These tokens can then be used to read, push, or overwrite images in other tenants' repositories. Deleting the attacker's namespace also deletes the corresponding robot accounts, which could disrupt systems that depend on them.

The response

Red Hat has released fixed builds for eight Quay 3 branches. The company recommends installing the fixed release as the only complete solution. For the first flaw, Red Hat also advises blocking the /keys endpoint at the proxy or load balancer for all but trusted internal components, searching audit logs for suspicious key rotation or deletion events, and removing any unrecognised service keys. For the second flaw, the company suggests limiting who can create organisations or namespaces with underscores, rotating robot tokens for affected namespaces, and monitoring for unusual activity. However, these are only partial mitigations until the update is applied. Red Hat says that turning off anonymous access does not protect against the main flaw, and that a planted key will still work after password resets or robot token changes.

Is there a fix?

Red Hat has released updates for all affected Quay versions. The upstream fix adds a missing rotation check, requires keys added through the key server to have an expiry date, and prevents callers from choosing a token's audience. The same update addresses the robot account search flaw. Quay.io, the hosted service, is already fixed. However, as of the report's publication, the CVE entry for CVE-2026-123710 still listed every Quay package as affected, with no fix, even though advisories had been published.

So far, only The Hacker News has reported these flaws, and there is no primary source from Red Hat or a government agency. Details may change as more information becomes available.

Why it matters

These flaws could let attackers steal sensitive data or inject malicious code into software supply chains. Even private, internal registries are at risk if not updated. The flaws are easy to exploit and affect widely used versions of Quay.

Key points

  1. All versions of Red Hat Quay 3 and 5 are affected by two critical flaws.
  2. Attackers can read and overwrite private container images, even without an account if anonymous access is enabled.
  3. Red Hat has released fixes, but partial mitigations are not enough.
  4. Robot account tokens can be stolen and used to inject malicious images.
  5. Immediate patching is essential for anyone running Quay registries.

Sources

  1. The Hacker News

    Reference

The day's cyber news in 5 minutes.

Cyberend Daily is one short email with the stories that matter, written so anyone can follow.

Includes our free beginner roadmap · No spam · Unsubscribe anytime