GCVE Workshop - 22 September 2026 (14:00-18:00), Luxembourg Before The Vulnopticon Conference - Registration

CVE-2026-86793 (GCVE-0-2026-86793)

Vulnerability from cvelistv5 – Published: 2026-09-11 11:35 – Updated: 2026-09-11 11:35
VLAI
Title
CVE-2026-86793
Summary
SGLang allows unauthenticated pickle deserialization through /update_weights_from_tensor when no auth keys are configured, and the SafeUnpickler policy can be bypassed because builtins.import and builtins.getattr are resolvable, enabling code execution via pickle REDUCE.
Severity
No CVSS data available.
CWE
Assigner
References
Impacted products
Vendor Product Version CPE status
SGLang SGLang Affected: 0 , ≤ 0.5.18 (custom)
guessed Create a notification for this product.
Show details on NVD website
To clipboard 

{
  "containers": {
    "cna": {
      "affected": [
        {
          "product": "SGLang",
          "vendor": "SGLang",
          "versions": [
            {
              "lessThanOrEqual": "0.5.18",
              "status": "affected",
              "version": "0",
              "versionType": "custom"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "SGLang allows unauthenticated pickle deserialization through /update_weights_from_tensor when no auth keys are configured, and the SafeUnpickler policy can be bypassed because builtins.import and builtins.getattr are resolvable, enabling code execution via pickle REDUCE."
        }
      ],
      "problemTypes": [
        {
          "descriptions": [
            {
              "description": "CWE-94 Improper Control of Generation of Code (\u0027Code Injection\u0027)",
              "lang": "en"
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-09-11T11:35:20.812Z",
        "orgId": "37e5125f-f79b-445b-8fad-9564f167944b",
        "shortName": "certcc"
      },
      "references": [
        {
          "url": "https://vicone.com/blog/cve-2026-86793-sglang-bypass-could-let-attackers-run-code-on-ai-servers/"
        },
        {
          "url": "https://github.com/sgl-project/sglang/tree/main"
        }
      ],
      "source": {
        "discovery": "UNKNOWN"
      },
      "title": "CVE-2026-86793",
      "x_generator": {
        "engine": "VINCE 3.0.45",
        "env": "prod",
        "origin": "https://cveawg.mitre.org/api/cve/CVE-2026-86793"
      }
    }
  },
  "cveMetadata": {
    "assignerOrgId": "37e5125f-f79b-445b-8fad-9564f167944b",
    "assignerShortName": "certcc",
    "cveId": "CVE-2026-86793",
    "datePublished": "2026-09-11T11:35:20.812Z",
    "dateReserved": "2026-09-08T11:56:03.107Z",
    "dateUpdated": "2026-09-11T11:35:20.812Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2",
  "vulnerability-lookup:meta": {
    "nvd": "{\"cve\":{\"id\":\"CVE-2026-86793\",\"sourceIdentifier\":\"cret@cert.org\",\"published\":\"2026-09-11T12:16:52.427\",\"lastModified\":\"2026-09-11T14:46:12.730\",\"vulnStatus\":\"Deferred\",\"cveTags\":[],\"descriptions\":[{\"lang\":\"en\",\"value\":\"SGLang allows unauthenticated pickle deserialization through /update_weights_from_tensor when no auth keys are configured, and the SafeUnpickler policy can be bypassed because builtins.import and builtins.getattr are resolvable, enabling code execution via pickle REDUCE.\"}],\"affected\":[{\"source\":\"cret@cert.org\",\"affectedData\":[{\"vendor\":\"SGLang\",\"product\":\"SGLang\",\"versions\":[{\"version\":\"0\",\"lessThanOrEqual\":\"0.5.18\",\"versionType\":\"custom\",\"status\":\"affected\"}]}]}],\"metrics\":{},\"references\":[{\"url\":\"https://github.com/sgl-project/sglang/tree/main\",\"source\":\"cret@cert.org\"},{\"url\":\"https://vicone.com/blog/cve-2026-86793-sglang-bypass-could-let-attackers-run-code-on-ai-servers/\",\"source\":\"cret@cert.org\"}]}}"
  }
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.





Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.


Detection rules are retrieved from Rulezet.