Skip to main content
Version: v1.4.x

Security Advisories

Security advisories supplement CVE reports, providing additional details regarding vulnerabilities and offering remediation steps.

Security Advisory 024 - eMudhra Certificate Authority Compromise (TLS Trust Store Exposure)​

  • Release Date: October 23, 2026
  • Last Updated: October 23, 2026
  • Severity: High
  • CVSS Score: 7.4 (CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:N)
  • CVE: CVE-2026-111943

Summary​

The private keys of eight eMudhra intermediate Certificate Authority (CA) certificates are recoverable by anyone because of a key-generation flaw from April 2025. The intermediates chain to two publicly trusted roots: emSign ECC Root CA - G3 and AAA Certificate Services, which cross-signs the same keys.

An attacker who holds one of these keys can issue a fraudulent X.509 certificate for any hostname. Any TLS client that trusts one of the roots and that an attacker can intercept accepts the forged certificate, which lets the attacker impersonate a TLS endpoint and compromise the confidentiality and integrity of the connection.

Affected Deployments​

Only self-hosted PaletteAI environments are affected. PaletteAI has no SaaS deployment, so no hosted environment is patched on your behalf.

The exposure is client-side only:

  • Not affected: Inbound and serving traffic. A PaletteAI endpoint presents its own certificate and consults no trust store.
  • Not affected: Traffic between PaletteAI components. Hub and spoke communication uses Open Cluster Management (OCM) with mutual TLS, and internal communication uses HTTPS with cert-manager-issued certificates, so the compromised roots are not consulted.
  • Affected: Outbound client connections that validate the other end against a system trust store that still trusts the compromised roots. This includes container image pulls and other outbound TLS connections from cluster hosts.

Impact​

Where an attacker holds a leaked key and an interception position, successful exploitation could allow them to:

  • Intercept and alter the hub-to-spoke control channel, including theft of the cluster authentication token.
  • Impersonate the management plane and inject control responses to a spoke cluster.
  • Intercept container image pulls, which can lead to a malicious image and node-level code execution.
  • Intercept other outbound TLS connections that rely on the system trust store, exposing credentials or data carried on them.

Unauthenticated exploitation of a PaletteAI endpoint is not possible through this issue. It requires both a leaked key and a network interception position.

  1. Upgrade self-hosted PaletteAI to 1.4.1 or later. All component images in 1.4.1 are rebuilt with Go 1.27.3, which rejects the compromised eMudhra intermediate keys. Refer to the Upgrade guide.
  2. Patch the host operating system on PaletteAI cluster hosts and spoke nodes. Refer to the OS Patching documentation.
  3. Monitor and audit outbound connections for anomalous activity.

References​