Two Critical Red Hat Quay Flaws Could Expose Private Images and Robot Tokens

Summary
Two critical Red Hat Quay flaws can expose private images and secrets, allow image tampering, or steal other tenants’ robot tokens. Red Hat lists affected versions and fixes, with token rotation and access restrictions suggested as interim steps.
Key points
- CVE-2026-123710 (CVSS 9.8) lets an attacker add a trusted signing key and use tokens to read, push, or overwrite images across repositories.
- The first flaw can expose secrets or credentials in private images; planted keys persist across restarts and may disrupt normal pushes and pulls.
- CVE-2026-110697 (CVSS 9.9) lets a Quay user who can create a namespace exploit unescaped SQL LIKE wildcards to obtain other tenants’ robot account tokens.
- Stolen robot tokens can enable image access or tampering; in OpenShift CI/CD pipelines, they may be used to inject malicious images, Red Hat says.
- Red Hat says Quay 3 and Quay 5 are affected and identifies fixed releases, but the article notes inconsistencies between CVE pages and advisories.
- Red Hat recommends updating; interim measures for the second flaw include restricting creation of underscore-containing namespaces, rotating potentially exposed robot tokens, and monitoring unusual image activity.
Article Details
- Vulnerability Types
- Unauthorized insertion of a trusted signing key through insufficient validation of key-rotation tokens
- Unescaped SQL LIKE wildcard in robot-account namespace lookup enables cross-tenant token disclosure
- Severity
- Critical (CVSS 9.8 for CVE-2026-123710; CVSS 9.9 for CVE-2026-110697)
- Affected Versions
- Quay 3 (all versions)
- Quay 5 (all versions)
- Exploitation Status
- unknown
- Exploit Availability
- unknown
- Patch Status
- partial
- Workarounds
- For CVE-2026-110697, limit who can create organizations or namespaces containing an underscore.
- Rotate robot tokens for namespaces that match an underscore-containing namespace.
- Monitor for pulls or pushes that do not fit normal robot-account use.
MITRE ATT&CK
T1078 · Valid AccountsStolen robot account tokens can be used to access other tenants' images.T1195.002 · Compromise Software Supply ChainThe article says stolen robot credentials can be used to inject malicious images into CI/CD pipelines when Quay is the integrated OpenShift registry.T1210 · Exploitation of Remote ServicesAn attacker who can reach a Quay registry exploits its flaws to access repositories and alter images.T1550.001 · Application Access TokenThe attacker adds a trusted signing key and uses tokens signed with that key to access repositories.T1565.001 · Stored Data ManipulationAn attacker can push or overwrite container images, and the changed images propagate to systems that later pull them.
CVE
CVE-2026-110697Red Hat · Red Hat Quay 3; Red Hat · Red Hat Quay 5A second critical flaw, CVE-2026-110697 (CVSS score: 9.9), was disclosed at the same time. It lets a Quay user who can create a namespace steal other tenants' robot account tokens and use them to read, push, orCVE-2026-123710Red Hat · Red Hat Quay 3; Red Hat · Red Hat Quay 5The flaw is tracked as CVE-2026-123710 (CVSS score: 9.8).
People
Vendors
Products
OpenShiftrobot accounts, which can cut off systems that rely on them. Where Quay is the integrated registry for OpenShift, stolen robot credentials can be used to inject malicious images into CI/CD pipelines, the company says.Red Hat QuayA critical flaw in Red Hat Quay allows an attacker who can access a Quay registry to read and overwrite private container images in every repository.
