CVE-2026-50563: Fission Container Executor PodSpec Injection Enables Kubernetes Node Escape

Vulnerability Overview

CVE-2026-50563 is a critical privilege escalation vulnerability (CVSS 9.9) in the Fission serverless framework, classified under CWE-269 (Improper Privilege Management) and CWE-284 (Improper Access Control). The root cause lies in Fission's container executor, which fails to sanitize or restrict user-supplied PodSpec fields before applying them to function pods. This allows an attacker to inject arbitrary Kubernetes PodSpec directives — effectively taking control of how the pod is configured at the platform level.

The impact is severe: successful exploitation results in a full container escape to the underlying Kubernetes node, granting an attacker root-level access to the host and lateral movement potential across the cluster.

Affected Software

• Product: Fission (github.com/fission/fission)
• Affected versions: All releases up to and including 1.23.0
• Vendor ecosystem: Go
• Patch status: No patch available as of time of writing — this is an unpatched vulnerability

Exploitation Status

A public proof-of-concept (PoC) exploit has been released. Disclosure occurred on 30 June 2026, making this a very recent, actively relevant threat. While it has not yet been confirmed as actively exploited in the wild or added to CISA's Known Exploited Vulnerabilities (KEV) catalogue, the availability of a public PoC dramatically lowers the barrier to exploitation. Security teams running Fission in any environment — particularly multi-tenant or internet-exposed clusters — should treat this as an urgent priority.

The attack surface is especially dangerous because exploitation does not require elevated privileges. An unauthenticated or low-privileged user can trigger the injection by supplying malicious PodSpec fields such as hostPID: true, hostNetwork: true, privileged container flags, hostPath volume mounts pointing to sensitive node paths, or custom service accounts with cluster-wide permissions.

Detection Coverage

Our detection logic targets the observable artefacts of PodSpec injection: anomalous pod creation events carrying dangerous security context fields, unexpected hostPID or hostNetwork flags on Fission-managed function pods, privileged container specifications, and suspicious hostPath volume mounts within the Fission namespace. Secondary signals include unusual service account bindings and pod admission events that deviate from the expected Fission executor baseline.

Full SIEM coverage ships for all major platforms:

• Microsoft Sentinel — KQL query over Kubernetes audit logs and ContainerInventory
• Splunk — SPL search against kube-audit and container telemetry sourcetypes
• Elastic — EQL sequence detecting pod creation with injected privileged fields
• IBM QRadar — AQL rule targeting Kubernetes API event flows
• Sumo Logic — CQL-based detection over ingested kube-apiserver audit events
• Google Chronicle — YARA-L rule correlating UDM Kubernetes resource modification events
• CrowdStrike — CQL detection via Falcon's Kubernetes Protection telemetry

Each query is tuned to minimise false positives by scoping to Fission executor namespaces and function pod naming conventions, while flagging any PodSpec field that would grant host-level access.

Recommendations

Until a patch is released, restrict access to the Fission API to trusted identities only, enforce Kubernetes admission controllers (OPA/Gatekeeper or Kyverno) to deny privileged PodSpec fields cluster-wide, and audit existing function deployments for signs of prior injection. Enable Kubernetes audit logging at the RequestResponse level to ensure full visibility into pod creation payloads.

View the full detection page — including all SIEM queries, mapped MITRE ATT&CK techniques, and atomic test guidance — at /detections/CVE-2026-50563.