Skip to content
Heist/ feed
Back to feed
High

vLLM: Public multimodal SHM handle forgery reaches worker-side `pickle.loads()` with insecure serialization disabled

Type: advisorySources: Official CVE List

vLLM is an inference and serving engine for large language models. Prior to 0.32.0, an authenticated caller can place forged shared-memory address and monotonic_id values in public multimodal features.kwargs_data. In scale-out deployments with shared-memory multimodal caching enabled, workers trust those values and pass attacker-selected bytes to pickle.loads(), allowing arbitrary Python execution even when insecure serialization is disabled. This issue is fixed in version 0.32.0.

Products
vllm
Affected versions
Not specified
Fixed versions
Not specified
CVEs
CVE-2026-115912
Confirmed fixed CVEs
Not specified
Updated in this feed
30 Oct 2026, 19:50 UTC · Revision 1

Source evidence

vLLM: Public multimodal SHM handle forgery reaches worker-side `pickle.loads()` with insecure serialization disabled

core-cve · Original source

Published 26 Oct 2026, 16:54 UTC

vLLM is an inference and serving engine for large language models. Prior to 0.32.0, an authenticated caller can place forged shared-memory address and monotonic_id values in public multimodal features.kwargs_data. In scale-out deployments with shared-memory multimodal caching enabled, workers trust those values and pass attacker-selected bytes to pickle.loads(), allowing arbitrary Python execution even when insecure serialization is disabled. This issue is fixed in version 0.32.0.

cve:CVE-2026-115912

vllm

Fixed versions: Not specified

Complete source record
{
  "id": "23094aca6f167f747c41c8807a3018dda71171350759fbe807cdda06b940cdb2",
  "url": "https://www.cve.org/CVERecord?id=CVE-2026-115912",
  "method": "structured",
  "payload": {
    "kind": "advisory",
    "title": "vLLM: Public multimodal SHM handle forgery reaches worker-side `pickle.loads()` with insecure serialization disabled",
    "summary": "vLLM is an inference and serving engine for large language models. Prior to 0.32.0, an authenticated caller can place forged shared-memory address and monotonic_id values in public multimodal features.kwargs_data. In scale-out deployments with shared-memory multimodal caching enabled, workers trust those values and pass attacker-selected bytes to pickle.loads(), allowing arbitrary Python execution even when insecure serialization is disabled. This issue is fixed in version 0.32.0.",
    "evidence": "cve:CVE-2026-115912",
    "products": [
      {
        "name": {
          "value": "vllm",
          "evidence": "/containers/cna/affected/0/product"
        },
        "prerequisites": [],
        "fixed_versions": [],
        "affected_versions": []
      }
    ],
    "severity": "high",
    "structured": {
      "id": "CVE-2026-115912",
      "aliases": [
        "CVE-2026-115912"
      ],
      "metrics": [
        {
          "cvssV3_1": {
            "scope": "UNCHANGED",
            "version": "3.1",
            "baseScore": 7.5,
            "attackVector": "NETWORK",
            "baseSeverity": "HIGH",
            "vectorString": "CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H",
            "integrityImpact": "HIGH",
            "userInteraction": "NONE",
            "attackComplexity": "HIGH",
            "availabilityImpact": "HIGH",
            "privilegesRequired": "LOW",
            "confidentialityImpact": "HIGH"
          }
        }
      ],
      "affected": [
        {
          "vendor": "vllm-project",
          "product": "vllm",
          "versions": [
            {
              "status": "affected",
              "version": "< 0.32.0"
            }
          ]
        }
      ],
      "modified": "2026-10-26T19:35:42.879Z",
      "provider": "cve",
      "references": [
        "https://github.com/vllm-project/vllm/security/advisories/GHSA-6757-hggm-hp2w",
        "https://github.com/vllm-project/vllm/pull/59357",
        "https://github.com/vllm-project/vllm/commit/3d5625cc3c2474994107047d7a6ecb8e4008f9b0",
        "https://github.com/vllm-project/vllm/releases/tag/v0.32.0"
      ],
      "provider_metadata": {
        "adp": [
          {
            "title": "CISA ADP Vulnrichment",
            "metrics": [
              {
                "other": {
                  "type": "ssvc",
                  "content": {
                    "id": "CVE-2026-115912",
                    "role": "CISA Coordinator",
                    "options": [
                      {
                        "Exploitation": "none"
                      },
                      {
                        "Automatable": "no"
                      },
                      {
                        "Technical Impact": "total"
                      }
                    ],
                    "version": "2.0.3",
                    "timestamp": "2026-10-26T19:31:18.550495Z"
                  }
                }
              }
            ],
            "providerMetadata": {
              "orgId": "134c704f-9b21-4f2e-91b3-4a467353bcc0",
              "shortName": "CISA-ADP",
              "dateUpdated": "2026-10-26T19:35:42.879Z"
            }
          }
        ],
        "cna": {
          "orgId": "a0819718-46f1-4df5-94e2-005712e83aaa",
          "shortName": "GitHub_M",
          "dateUpdated": "2026-10-26T16:54:00.643Z"
        }
      }
    },
    "related_cves": [
      {
        "value": "CVE-2026-115912",
        "evidence": "/cveMetadata/cveId"
      }
    ],
    "security_relevant": true,
    "severity_evidence": "/containers/cna/metrics"
  },
  "source_id": "core-cve",
  "document_id": 338254,
  "published_at": "2026-10-26T16:54:00.643+00:00"
}