OSAC operator CVE-2026-62284

Exploitation requires the deployment to run in remote or auto VM cluster mode (--vm-cluster-mode; auto is the default); deployments using local mode exclusively are not affected because they never read kubeconfig Secrets. The attacker must be able to create or modify a Secret carrying the label osac.openshift.io/remote-cluster-kubeconfig: "true" in the namespace where the osac-console-proxy-secret-reader RoleBinding grants the proxy Secret read access. The limiting factor is this write privilege (reflected as PR:H): it is not reachable by an unauthenticated or unprivileged outsider, so realistic risk depends on who holds Secret create/modify rights in that namespace. The RCE requires the exec (or auth-provider) stanza in the kubeconfig; the secondary SSRF requires only an attacker-chosen clusters[].server value and remains exploitable after the exec-path patch.

This is a genuine, well-substantiated code-injection flaw (CWE-94) rather than an inflated high-CVSS issue, but its real-world urgency is tempered by a meaningful privilege prerequisite. The supplied CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:H/UI:N/VC:H/VI:H/VA:H) correctly reflects high impact (full RCE inside the console-proxy pod as its service account) reachable over the network with low complexity, but PR:H is the decisive signal: the attacker must already be able to create or modify Secrets in the specific namespace where the osac-console-proxy-secret-reader RoleBinding grants read access. The vuln.today decision label is 'this_week' (not act_now/emergency), consistent with a serious-but-gated issue. No EPSS score, no CISA KEV listing, and no confirmed public exploit were provided, so there is no evidence of active or widespread exploitation - the risk is latent and dependent on a tenant/RBAC boundary being crossable. One important nuance: the exec credential-plugin RCE path is patched in commit 7e4222b, but the advisory and description both state that patch does NOT remediate the server-side request forgery path via attacker-controlled clusters[].server, so even patched deployments retain residual SSRF exposure from the proxy pod.

An attacker (or compromised workload) with permission to create or modify Secrets in the namespace watched by the console-proxy crafts a kubeconfig whose users[].user.exec stanza specifies an arbitrary command, stores it in a Secret labeled osac.openshift.io/remote-cluster-kubeconfig: "true", and waits for the proxy to resolve it. When the proxy builds the upstream WebSocket transport via rest.TransportFor, client-go invokes the exec plugin and runs the attacker's process as the osac-console-proxy service account inside the pod. No public exploit was identified at time of analysis, though the GHSA advisory and the PR #321 test fixture fully document the triggering kubeconfig.