Skip to main content
CVE Vulnerability Database
Vulnerability Database/CVE-2026-18694

CVE-2026-18694: MongoDB Server DoS Vulnerability

CVE-2026-18694 is a denial of service flaw in MongoDB Server's geospatial query processing that allows authenticated users to crash the server through malformed geometry data. This article covers technical details, impact, and mitigations.

Published:

CVE-2026-18694 Overview

CVE-2026-18694 is an out-of-bounds read vulnerability [CWE-125] in MongoDB Server's geospatial query processing. An authenticated user with write privileges can store malformed geometry data that later triggers memory access beyond intended bounds during query execution. The condition results in a server crash and may expose a limited amount of server process memory.

The issue affects MongoDB Server deployments that accept geospatial data from users with write access. Exploitation requires only low privileges over the network and no user interaction.

Critical Impact

Authenticated attackers with write access can crash the MongoDB server and potentially read fragments of process memory by injecting malformed geometry data used in subsequent geospatial queries.

Affected Products

  • MongoDB Server (see the MongoDB Server Issue Tracker for affected versions)
  • Deployments accepting geospatial (2d, 2dsphere) data from write-privileged users
  • Applications that execute geospatial queries against user-supplied geometry

Discovery Timeline

  • 2026-08-11 - CVE-2026-18694 published to the National Vulnerability Database (NVD)
  • 2026-08-11 - Last updated in NVD database

Technical Details for CVE-2026-18694

Vulnerability Analysis

The vulnerability resides in MongoDB Server's geospatial query processing pipeline. The server accepts geometry data on write operations without fully validating structural constraints. When a subsequent geospatial query, such as $geoIntersects or $geoWithin, processes the stored document, the query engine reads memory outside the intended bounds of the geometry buffer.

The out-of-bounds read produces two outcomes. First, the server process can crash, creating a denial-of-service condition for all clients connected to that node. Second, the returned data may include bytes from adjacent memory regions, exposing a limited portion of MongoDB's process memory.

Because the flaw is triggered by stored data rather than the write itself, exploitation is time-shifted from ingestion to query execution. This delayed trigger complicates incident correlation and can affect replica set secondaries that replay the same operations.

Root Cause

The root cause is missing validation of geometry data before it is persisted, combined with the assumption that stored geometries are well-formed at query time [CWE-125]. When the query processor iterates coordinate arrays or vertex structures, boundary checks are insufficient to prevent reads past the allocated region.

Attack Vector

An authenticated user with write privileges submits a document containing malformed GeoJSON or legacy coordinate geometry to a collection with a geospatial index. When any client, including automated aggregation jobs, later queries that data with a geospatial operator, the server dereferences memory outside the geometry buffer. Refer to the MongoDB Server Issue Tracker (SERVER-130188) for further technical details.

Detection Methods for CVE-2026-18694

Indicators of Compromise

  • Unexpected mongod process crashes or restarts correlated with geospatial query workloads
  • Segmentation fault or Fatal Assertion entries in MongoDB logs referencing geometry, 2d, or 2dsphere code paths
  • Writes containing unusually structured GeoJSON documents from accounts with write privileges
  • Replica set secondaries crashing shortly after replicating specific insert or update operations

Detection Strategies

  • Parse MongoDB audit and diagnostic logs for repeated crashes tied to geospatial operators such as $geoWithin, $geoIntersects, or $near
  • Alert on write operations to collections with geospatial indexes that contain malformed or oversized coordinate arrays
  • Correlate crash timestamps across primary and secondary nodes to identify replicated malicious documents

Monitoring Recommendations

  • Baseline mongod uptime and crash frequency, alerting on statistically significant deviations
  • Track authenticated write activity per role and flag accounts writing to geo-indexed collections outside normal patterns
  • Forward MongoDB logs to a central analytics platform to correlate crashes with the originating write operations

How to Mitigate CVE-2026-18694

Immediate Actions Required

  • Apply the MongoDB Server security update referenced in SERVER-130188 as soon as fixed builds are available for your release train
  • Audit and restrict write privileges on collections containing geospatial indexes to trusted service accounts
  • Review recent writes to geo-indexed collections for malformed geometry and quarantine suspect documents
  • Enable database auditing to capture write operations against geospatial collections

Patch Information

MongoDB tracks the fix under issue SERVER-130188. Consult the MongoDB Server Issue Tracker for the list of patched server versions and upgrade guidance. Apply the vendor-supplied update to primaries, secondaries, and arbiters, then validate replica set health after each node upgrade.

Workarounds

  • Remove or disable geospatial indexes on collections that accept untrusted input until patching is complete
  • Enforce application-layer validation of GeoJSON structure, coordinate counts, and ranges before insert or update
  • Revoke write privileges from accounts that do not require geospatial write access using MongoDB role-based access control
  • Deploy a query proxy or driver-level filter that rejects geospatial operators from low-trust clients
bash
# Example: revoke write access on a geo-indexed collection using mongosh
use admin
db.revokeRolesFromUser("appuser", [
  { role: "readWrite", db: "geoData" }
])
db.grantRolesToUser("appuser", [
  { role: "read", db: "geoData" }
])

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Default Legacy - Prefooter | Experience the World’s Most Advanced Cybersecurity Platform

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.