Skip to main content
Vulnerability Database/CVE-2026-105447

CVE-2026-105447: Quay Privilege Escalation Vulnerability

CVE-2026-105447 is a privilege escalation vulnerability in Quay that allows read-only administrators to obtain repository write tokens through build trigger APIs, enabling unauthorized container image uploads to private repositories.

Published:

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

CVE-2026-105447 Overview

CVE-2026-105447 is an authorization flaw in Red Hat Quay container registry. The application exposes build trigger configuration details containing repository write tokens to global read-only administrative users. An authenticated user with read-only privileges can query the build trigger API and retrieve delegate tokens intended for higher-privileged accounts.

Attackers abusing these tokens bypass read-only restrictions and push arbitrary container images into private repositories. This enables privilege escalation within the registry and potential supply chain compromise for downstream consumers of those images. The flaw is classified under CWE-863: Incorrect Authorization.

Critical Impact

Read-only administrators can harvest repository write tokens from build trigger responses and push malicious container images into private Quay repositories.

Affected Products

  • Red Hat Quay container registry
  • Quay deployments exposing the build trigger API to read-only administrative users
  • Downstream distributions and operators embedding affected Quay versions

Discovery Timeline

  • 2026-10-05 - CVE-2026-105447 published to the National Vulnerability Database
  • 2026-10-06 - Last updated in NVD database

Technical Details for CVE-2026-105447

Vulnerability Analysis

Quay's build trigger subsystem stores delegate credentials that allow automated builds to push images into target repositories. When a global read-only administrator queries the trigger configuration endpoint, the response includes these delegate tokens in full. The access control decision does not distinguish between inspecting trigger metadata and viewing embedded secrets.

An attacker holding read-only administrative credentials can enumerate build triggers across tenants and extract write-capable tokens. Those tokens then authenticate container push operations against private repositories, effectively granting write access the account was never meant to hold. The outcome is image tampering, backdoored builds, and compromised artifacts delivered to any pipeline consuming the registry.

Root Cause

The root cause is an incorrect authorization check [CWE-863] in the build trigger API handler. The endpoint returns sensitive trigger configuration data, including delegate write tokens, without filtering fields based on the caller's effective permissions. Read-only roles were expected to see metadata only, but the serializer emits the complete trigger object.

Attack Vector

Exploitation requires an authenticated session with high privileges, specifically a global read-only administrative role. The attacker issues standard HTTP requests to the build trigger API, parses the JSON response, and extracts the delegate token. No user interaction is required, and the attack is executed over the network against the Quay API surface. See the Red Hat CVE-2026-105447 Advisory and Red Hat Bug Report #2545916 for vendor technical details.

Detection Methods for CVE-2026-105447

Indicators of Compromise

  • Unexpected GET requests to Quay build trigger configuration endpoints originating from read-only administrator accounts.
  • Container image push events authenticated with delegate tokens from accounts that do not normally initiate builds.
  • New or modified image tags in private repositories without a corresponding CI/CD pipeline run.

Detection Strategies

  • Audit Quay API access logs for read-only accounts enumerating /api/v1/repository/*/trigger/ endpoints.
  • Correlate token issuance events with subsequent docker push or OCI push activity from unusual source addresses.
  • Alert on image digest changes in private repositories that lack an associated signed build attestation.

Monitoring Recommendations

  • Forward Quay application and audit logs to a centralized analytics platform for retention and correlation.
  • Baseline normal build trigger API consumers and alert on deviations in caller identity or query frequency.
  • Monitor registry write operations against expected CI/CD service account identities and reject outliers.

How to Mitigate CVE-2026-105447

Immediate Actions Required

  • Review and rotate any delegate tokens associated with existing build triggers in Quay.
  • Audit assignment of the global read-only administrator role and remove it from accounts that do not require it.
  • Inspect recent image pushes to private repositories for unauthorized or unexpected artifacts.

Patch Information

Refer to the Red Hat CVE-2026-105447 Advisory for the authoritative list of fixed packages and errata. Apply vendor-supplied updates to all Quay instances once fixed versions are available. Track remediation status through Red Hat Bug Report #2545916.

Workarounds

  • Restrict access to the build trigger API at the network or reverse proxy layer until a patch is applied.
  • Revoke the global read-only administrator role from non-essential users as a short-term containment measure.
  • Rotate repository credentials and OAuth application secrets referenced by existing build triggers.
bash
# Example: restrict access to the build trigger API at the reverse proxy
# Replace with your ingress or load balancer configuration syntax
location ~ ^/api/v1/repository/.*/trigger/ {
    allow 10.0.0.0/8;      # CI/CD management network
    deny  all;
    proxy_pass http://quay-backend;
}

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.