criticalZero-Day

OpenShift Console Devfile API Unauthenticated SSRF and DoS

First seen Sep 20, 2026 · Updated Sep 20, 2026 · CVSS 9.3

openshiftkubernetesssrfdosunauthenticatedconsoledevfilecloud-nativeagent-relevant

An unauthenticated flaw in the OpenShift console's devfile API endpoints allows remote attackers to trigger SSRF requests from the console pod to internal cluster services, potentially exposing internal responses. The same endpoints are also vulnerable to a memory exhaustion DoS via unbounded, content-length-less large requests. The vulnerability requires no authentication, making it highly exploitable at scale against exposed OpenShift consoles.

Technical Analysis

CVE-2026-75885 (CVSS 9.3) affects the `/api/devfile/` and `/api/devfile/samples/` endpoints of the OpenShift web console, which process attacker-supplied devfile payloads without authentication or adequate validation. By crafting malicious devfile URLs, an attacker can force the console pod to issue SSRF requests against internal-only services (e.g., cloud metadata endpoints, internal APIs, or in-cluster microservices) and receive partial reflected responses, enabling internal network reconnaissance and potential credential/secret disclosure. A separate vector allows attackers to send large payloads without a Content-Length header, causing unbounded memory allocation in the console pod and resulting in denial of service or pod eviction/crash. Because many organizations deploy AI agent orchestration, RAG pipelines, and MLOps tooling on OpenShift/Kubernetes clusters, SSRF-driven access to internal service endpoints could expose vector database credentials, model-serving API keys, or agent tool-integration secrets stored in cluster-internal services, and the DoS could disrupt agent-dependent workloads relying on the console for cluster management. This makes the flaw agent-relevant wherever OpenShift hosts AI agent infrastructure.

Affected Systems

Red Hat OpenShift Container Platform web console (all versions exposing the `/api/devfile/` and `/api/devfile/samples/` endpoints prior to patched release); affects any OpenShift cluster with the console component enabled and network-reachable, including managed OpenShift offerings (ROSA, ARO, OpenShift Dedicated) unless mitigated.

Indicators of Compromise

  • Anomalous outbound requests from console pod to internal-only IP ranges (e.g., 169.254.169.254 cloud metadata service)
  • Repeated unauthenticated POST/GET requests to /api/devfile/ and /api/devfile/samples/ endpoints
  • Requests lacking Content-Length header with large payload bodies
  • Unexpected memory growth or OOM-kill events on console pod
  • Suspicious devfile URLs referencing internal hostnames or RFC1918 addresses in access logs

Remediation Steps

  1. 1

    Apply vendor patch

    Upgrade OpenShift Container Platform to the version containing the fix for CVE-2026-75885 as soon as it is released by Red Hat; monitor Red Hat security advisories for the patched release.

  2. 2

    Restrict network exposure

    Limit access to the OpenShift console and its API endpoints to trusted networks via ingress/network policies, and disable public exposure of the console where not required.

  3. 3

    Implement egress filtering

    Apply network policies or a service mesh to restrict console pod egress to only necessary internal endpoints, blocking access to cloud metadata services and sensitive internal APIs.

  4. 4

    Add request validation

    Deploy a WAF or reverse proxy rule to reject devfile requests missing Content-Length headers or exceeding reasonable payload size limits.

  5. 5

    Monitor and alert

    Enable logging and alerting on console pod resource usage spikes and unauthenticated access attempts to devfile endpoints.

  6. 6

    Audit exposed secrets

    Review and rotate any credentials, API keys, or secrets accessible to internal services that could be reached via SSRF from the console pod, including those used by AI agent or MLOps tooling.

CVE / Advisory IDs

CVE-2026-75885

Industries Most Exposed

technologycloud servicesfinancial serviceshealthcaregovernmenttelecommunicationsretailmanufacturing

Sources

Respond to this threat

Pro subscribers get a full AI-generated incident-response playbook for this threat — detection, containment, eradication, and recovery steps — plus an unlimited AI Threat Advisor for questions about your environment.