CVE-2026-84196

Kyverno, a widely used open-source policy engine for Kubernetes, contains a server-side request forgery (SSRF) vulnerability in the apiCall service URL field (apiCall.service.url) in all releases before 1.18.0. An authenticated user who can create or modify Kyverno policies can inject user-controlled input through variable substitution in that field, causing Kyverno to issue arbitrary HTTP requests to attacker-chosen targets, including internal cluster services, cloud metadata endpoints (such as 169.254.169.254), and loopback addresses. Because the HTTP response is reflected back in admission error messages, the SSRF is non-blind, allowing attackers to read internal service responses and exfiltrate the data through the cluster's admission API. Any Kubernetes cluster running a vulnerable Kyverno release is affected, with exposure limited to users or compromised accounts holding permissions to create or update Kyverno policy resources. There is no evidence of exploitation in the wild: no public proof-of-concept is known, the flaw is not in CISA's KEV catalog, and EPSS estimates roughly a 0.3% probability of exploitation in the next 30 days.

What to do: Upgrade to Kyverno 1.18.0 or later. As interim mitigation, restrict RBAC permissions for creating or updating Kyverno Policy/ClusterPolicy resources to trusted administrators and audit existing apiCall-based policies for variable substitutions sourced from user-controlled fields. Consider restricting Kyverno pod egress at the network level (blocking the cloud metadata endpoint 169.254.169.254 and loopback ranges) to limit what the SSRF can reach.