Skip to content
KitploitKITPLOIT
ToolsExploitsBlog
Log in
Submit
ToolsExploitsBlog
Submit

Hacking, PenTest, and Cybersecurity Tools for Your Security Arsenal!

Kitploit is a directory of hacking, cybersecurity, and pentesting tools. Discover the latest project updates to find vulnerabilities, analyze systems, automate testing, and strengthen your security.

FeedsContactPrivacy© 2026 Kitploit

Tool Directory

Categories

View all categories
Loading categories
CVE-2026-130587 — PoC lab for CVE-2026-130587: demonstrates OpenFunction <=1.2.0 privilege escalation where a tenant-supplied Serving PodSpec yields privileged host-namespace pods. | Kitploit
Tools/GitHubGitHub/googlesky/cve-2026-130587
Cloud Infrastructure SecurityPrivilege EscalationContainer SecurityExploitationPenetration TestingCloud SecurityLabs & Practice
GitHubgooglesky/cve-2026-130587

CVE-2026-130587

PoC lab for CVE-2026-130587: demonstrates OpenFunction <=1.2.0 privilege escalation where a tenant-supplied Serving PodSpec yields privileged host-namespace pods.

View Repository
1 day agoNot yet reviewed
Share

CVE-2026-130587 — OpenFunction: tenant-controlled PodSpec in Function/Serving is copied into operator workloads (privilege escalation)

License: MIT Research status: validated

CVECVE-2026-130587
ComponentOpenFunction — pkg/core/serving/openfuncasync/servingrun.go::generateWorkload()
AffectedOpenFunction through 1.2.0
Fixed inno vendor fix (upstream issue #532 OPEN); the lab's fixed side carries a documented synthetic boundary guard (sha256 in evidence/versions.txt)
CWECWE-269 (Improper Privilege Management), CWE-266 (Incorrect Privilege Assignment)
ImpactA namespace tenant who can create Function/Serving resources — but has no Pod-create or cluster-admin rights — makes the operator schedule a privileged pod with host namespaces and a hostPath mount of the node root filesystem: escape from the namespace to node access
VectorKubernetes API — Serving.spec.template (a full corev1.PodSpec)
Triggerone CR whose spec.template sets privileged: true, hostPID/hostNetwork/hostIPC, and a hostPath: / volume

Serving.Spec.Template is a *corev1.PodSpec supplied by the tenant. The operator takes it verbatim:

// pkg/core/serving/openfuncasync/servingrun.go @ v1.2.0 (~L320)
spec := s.Spec.Template
if spec == nil {
    spec = &corev1.PodSpec{}
}
...
// `spec` becomes the PodTemplateSpec of the generated Deployment /
// StatefulSet / Job, created with the operator's own credentials

No field filtering, Pod Security admission, or ValidatingAdmissionPolicy is applied by the project. This lab builds a Go driver against the real v1.2.0 controller module and calls its actual generateWorkload() method, so the evidence is the operator's own output — not a simulation of the CRD shape.


1. Root cause

Kubernetes RBAC separates create functions/servings from create pods. The operator bridges that gap unsafely: the tenant-authored PodSpec is passed through to the generated workload. Security-relevant fields that a tenant could never set on a Pod directly — host namespaces, hostPath volumes, privileged containers, host ports, node selectors — are therefore granted to the tenant through the operator's service account. The resulting pod can mount the node root and read kubelet credentials, i.e. escalate from a single namespace to node/cluster level.

2. Reproduction environment

serviceimagerole
of-vulncve-2026-130587:vulnGo driver against unmodified OpenFunction v1.2.0
of-fixedcve-2026-130587:fixedsame module + synthetic PodSpec guard (issue #532 has no vendor fix)
poccve-2026-130587-poc:localstdlib-only Python driver

The HTTP service models the operator boundary: POST /generate accepts a Serving CR, runs the real workload generator as a child process, and reports the host-security fields of the resulting workload. The harness uses a fake controller-runtime client and the same-package bridge only to reach the unexported generator; the generator itself is upstream code on both sides.

3. Running the environment

./run-lab.sh build     # Go builds both trees (v1.2.0 and v1.2.0 + guard)
./run-lab.sh up        # cve130587lab network: of-vuln / of-fixed
./run-lab.sh sweep     # check / exploit / probe / mitigation on both sides
./run-lab.sh down

4. The PoC

A standalone full-chain exploit script ships at exploit/exploit.py (python3 exploit/exploit.py --host=<target>; exit 1 on confirmed exploitation, 0 on a safe target).

poc/poc.py (stdlib only) constructs Serving CRs with a hostile spec.template:

  • check — benign Serving vs hostile Serving; exit 1 if host/escape fields appear in the generated workload, 0 if refused, 3 no worker, 2 unreachable.
  • exploit — narrated chain from "tenant has only create-Serving" to "privileged host-namespace pod scheduled by the operator".
  • probe — per-field matrix: privileged, hostPID, hostNetwork, hostIPC, hostPath, plus a benign control.
  • mitigation — the synthetic guard on the fixed side; recommends Pod Security restricted plus a ValidatingAdmissionPolicy on Serving.spec.template until upstream ships a fix.

5. Expected results

caseof-vuln (v1.2.0)of-fixed (synthetic guard)
benign Servingworkload generated, no escape fieldsworkload generated
hostile ServinghostPID/hostNetwork/hostIPC true, hostPath: /, privileged containerrejected: "host namespaces are not allowed"
probe privileged / hostPID / hostNetwork / hostIPC / hostPathall five reach the workloadall five rejected with distinct errors

Captured runs: evidence/check-*-run.txt, evidence/exploit-*-run.txt, evidence/probe-*-run.txt, evidence/mitigation-*-run.txt, evidence/versions.txt, evidence/0day-sweep.txt.

6. Environment preconditions

  • Podman (or Docker); containers bounded on the isolated cve130587lab network.
  • Outbound HTTPS at image build (GitHub: OpenFunction tag; Go modules).
  • Serving.spec.triggers must be present ({}) or the controller panics in GetDaprServiceEnabled — a harness detail documented in evidence/versions.txt.
  • The Go binary is built with CGO_ENABLED=0 for the alpine runner.

7. Additional findings (0-day sweep on the fixed line)

The per-field matrix (evidence/0day-sweep.txt) confirms the synthetic guard intercepts all five escaping fields individually while leaving the benign path intact. No bypass form was found in the generator: the check runs after the PodSpec is read from the CR and before any workload object is built. Because upstream issue #532 remains open and no fixed release exists, the vendor-side result is pending; deployments should enforce Pod Security restricted on the operator namespace and validate Serving.spec.template with a ValidatingAdmissionPolicy now.

8. Repository layout

README.md             this write-up
run-lab.sh            environment driver (build / up / sweep / down)
Makefile              convenience targets
exploit/exploit.py    standalone full-chain exploit
poc/poc.py            the PoC (check / exploit / probe / mitigation)
poc/Dockerfile        runner image (python:3.12-alpine)
lab/lab.Dockerfile    multi-stage build of both controller trees
lab/patches/validate-podspec.patch   the synthetic fixed-side guard
lab/harness/          same-package bridge + Go driver
lab/service.py        HTTP boundary service (POST /generate)
evidence/             captured runs of every mode on both sides
TRACKING.md           excluded (working record)

9. References

  • CVE record — https://www.cve.org/CVERecord?id=CVE-2026-130587
  • Upstream issue (open) — https://github.com/OpenFunction/OpenFunction/issues/532
  • Vulnerable code @ v1.2.0 — https://github.com/OpenFunction/OpenFunction/blob/v1.2.0/pkg/core/serving/openfuncasync/servingrun.go#L320
  • Serving API type — https://github.com/OpenFunction/OpenFunction/blob/v1.2.0/apis/core/v1beta2/serving_types.go#L225
  • VulnCheck advisory — https://www.vulncheck.com/advisories/openfunction-through-1.2.0-privilege-escalation-via-unrestricted-podspec-template
Download Tool

Most Popular

View all →

Discover the most used tools by our community.

  • 1

    OsintHunter

    letox4/osinthunter

    🎯 The ultimate OSINT toolkit: investigation graph, confidence scoring, alias families, map, timeline and reports. 49 modules, 400+ sites, no API keys required.

    936
  • 2

    ghostlock-app

    yukonga/ghostlock-app

    GhostLock One-Tap Execution App (CVE-2026-43499)

    555
  • 3

    redforge

    cybersecurity91/redforge

    Browser-side red team toolkit with 102 offline tools for recon, payload generation, credential attacks, crypto, encoding and engagement reporting.

    353
  • 4

    ARTEX

    baarai/artex

    LLM multi-agent autonomous penetration testing platform with planner/worker agents, asset and exploration graphs, MITM traffic recording, and vulnerability retesting.

    348
  • 5

    looking-glass

    anonghosty/looking-glass

    CSIRT platform that ingests Shadowserver reports into MongoDB and replays them as interactive maps, analysis, findings, advisories and executive dashboards.

    336
  • 6

    CyberStrikeAI

    ed1s0nz/cyberstrikeai

    The system of action for AI-native cybersecurity—where intent becomes governed execution, evidence becomes operational memory, and every operation improves the next.

    288

Explore all tools

Browse our collection of tools

View all tools →