← back to feed
CVE-2026-111768

CVE-2026-111768

waiting for further signalsZero needs more public information before it can validate it.

GHSA-46jr-4wx8-v453 ↗

no check availablenot proven exploitable
awaiting cvepublic pocExploit code has been published by someone. Source signals can also carry a PoC (Proof of Concept).kevAttacks confirmed in the real world.
npm:@modelcontextprotocol/sdk >= 1.3.0, < 1.32.0; >= 2.0.0, < 2.3.0
fixed in 1.32.0, 2.3.0
CWE-770 Allocation of Resources Without Limits or Throttling
ETK-2026-1685Zero's internal tracking ID, independent of the CVE number., permanent identifier
pkg:npm/@modelcontextprotocol/sdk
pkg:npm/@modelcontextprotocol/server
cvss7.5cna vendorCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
6d 03h 56msince last run
14mAI analysisAnalysis: how long the AI spent reversing this signal.
timeline
0 = trigger signal
waiting for further signals−0.5dtrigger signal +0d
+2d
+4d
+6d

description

@modelcontextprotocol/sdk 1.3.0 through 1.31.0 and @modelcontextprotocol/server 2.0.0 through 2.2.0 do not limit the number of values validated in HTTP tool inputs. An unauthenticated remote client may cause excessive memory use and terminate an exposed server when untrusted clients can reach its HTTP interface; stdio-only deployments are not affected.

mitigation

Upgrade @modelcontextprotocol/sdk to 1.32.0 or later, or @modelcontextprotocol/server to 2.3.0 or later, and configure maxToolInputElements for the largest legitimate tool input. Also set a suitably small maxRequestBodySize on the transport, handler, or HTTP body parser; if an upgrade is not yet possible, enforce a request-body limit or place a limiting reverse proxy in front of the service.

fixed in 1.32.0, 2.3.0

check
no
exploit
no
automationDeterministic: a check that runs on its own, the same request on every target. Contextual: an agentic technique that needs to be applied per target, for bugs that depend on contextual flows.
contextual
auth requiredWhether an attacker needs an account first, based on the CVSS vector's Privileges Required.
no
EPSSEstimate of the chance this CVE is exploited in the next 30 days. Refreshed daily.
–
agent run time
14m
cve.org record, t = 0The moment cve.org publishes the CVE in full.
not published yet
Δ to public pocExploit code has been published by someone. Source signals can also carry a PoC (Proof of Concept).
–
Δ to kevAttacks confirmed in the real world.
–

outcome

Contextual: an exploit technique exists, but no deterministic check could cover this one.

source signals

Public signals that triggered or updated the pipeline. These are inputs for the deltas above.

  1. 2026-10-15 13:10Z

this entry in the feed

GET /api/v1/feed/CVE-2026-111768

{
  "etk_id": "ETK-2026-1685",
  "cve_id": "CVE-2026-111768",
  "display_id": "CVE-2026-111768",
  "ghsa_id": "GHSA-46jr-4wx8-v453",
  "euvd_id": null,
  "status": "published",
  "anchor_state": "awaiting_cve",
  "title": "Unbounded input validation in MCP SDK HTTP servers",
  "summary": "@modelcontextprotocol/sdk 1.3.0 through 1.31.0 and @modelcontextprotocol/server 2.0.0 through 2.2.0 do not limit the number of values validated in HTTP tool inputs. An unauthenticated remote client may cause excessive memory use and terminate an exposed server when untrusted clients can reach its HTTP interface; stdio-only deployments are not affected.",
  "mitigation": "Upgrade @modelcontextprotocol/sdk to 1.32.0 or later, or @modelcontextprotocol/server to 2.3.0 or later, and configure maxToolInputElements for the largest legitimate tool input. Also set a suitably small maxRequestBodySize on the transport, handler, or HTTP body parser; if an upgrade is not yet possible, enforce a request-body limit or place a limiting reverse proxy in front of the service.",
  "vendor": "npm",
  "product": "npm:@modelcontextprotocol/sdk",
  "affected": [
    {
      "affected_notes": [],
      "affected_ranges": [
        ">= 1.3.0, < 1.32.0"
      ],
      "ecosystem": "npm",
      "fixed_versions": [
        "1.32.0"
      ],
      "package_name": "@modelcontextprotocol/sdk",
      "purl": "pkg:npm/@modelcontextprotocol/sdk"
    },
    {
      "affected_notes": [],
      "affected_ranges": [
        ">= 2.0.0, < 2.3.0"
      ],
      "ecosystem": "npm",
      "fixed_versions": [
        "2.3.0"
      ],
      "package_name": "@modelcontextprotocol/server",
      "purl": "pkg:npm/@modelcontextprotocol/server"
    }
  ],
  "cwes": [
    770
  ],
  "cvss_score": 7.5,
  "cvss_src": "cna_vendor",
  "cvss_vector": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
  "epss_score": null,
  "epss_percentile": null,
  "epss_asof": null,
  "automation": "contextual",
  "check": false,
  "check_confidence_basis": null,
  "exploit": false,
  "pipeline_minutes": 14.02534395,
  "t0_at": null,
  "signals": {
    "trigger_signal": {
      "at": "2026-10-15T13:10:32+00:00",
      "days": -5.691980859074074,
      "is_lower_bound": true,
      "measured_at": "2026-10-21T05:46:59.169214+00:00",
      "url": "https://github.com/modelcontextprotocol/typescript-sdk/security/advisories/GHSA-46jr-4wx8-v453"
    }
  }
}

published 2026-10-15 14:44Z
last change 2026-10-21 05:31Z

Need the technical details behind an entry?

Every entry here shows what we attempted and when. The technical details, the generated exploit that fired it and the evidence behind the verdict are not public. Tell us about your use case and we’ll get back to you about extended API access.

Protected by reCAPTCHA; the Google Privacy Policy and Terms of Service apply.