CVE-2026-55253: LangChain MongoDB NoSQL injection in list/search methods

CVE-2026-55253 · Severity: high · CVSS 7.7 · Published 2026-09-14

LangChain's MongoDB checkpoint and store libraries use MongoDB to persist application state across conversations. These libraries fail to properly validate user input in filter parameters, allowing authenticated users to inject MongoDB query operators (like $ne or $regex) directly into database queries. An attacker can bypass tenant isolation and read another tenant's confidential checkpoint data without permission.

Vendors: LangChain, PyPI.

Quick answers

Is CVE-2026-55253 exploited in the wild?
No. Junglewise Threat Intelligence has no report of CVE-2026-55253 being exploited in the wild.
How severe is CVE-2026-55253?
High severity, with a CVSS score of 7.7, published 14 September 2026.
Which products and versions are affected?
langgraph-checkpoint-mongodb (PyPI) ECOSYSTEM:introduced=0,fixed=0.3.0; langgraph-store-mongodb (PyPI) < 0.4.0, ECOSYSTEM:introduced=0,fixed=0.4.0.
Is there a fix for CVE-2026-55253?
Yes. langgraph-checkpoint-mongodb (PyPI): upgrade to 0.3.0 or later; langgraph-store-mongodb (PyPI): upgrade to 0.4.0 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.
How is CVE-2026-55253 attacked?
It is attacked over the network, in a software library.
Am I affected by CVE-2026-55253?
Check whether you run LangChain LangGraph Checkpoint MongoDB, langgraph-checkpoint-mongodb (PyPI) or langgraph-store-mongodb (PyPI) at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

How to fix

  • langgraph-checkpoint-mongodb (PyPI): upgrade to 0.3.0 or later
    pip install "langgraph-checkpoint-mongodb>=0.3.0"
  • langgraph-store-mongodb (PyPI): upgrade to 0.4.0 or later
    pip install "langgraph-store-mongodb>=0.4.0"

Technical details

A NoSQL injection vulnerability exists in langgraph-checkpoint-mongodb (MongoDBSaver.list() and MongoDBSaver.alist()) and langgraph-store-mongodb (MongoDBStore.search()) libraries. The vulnerability arises because the filter parameter passed to these methods is incorporated directly into MongoDB query objects without sanitizing or rejecting MongoDB operator keys (prefixed with $). An authenticated caller who controls the filter input can embed query operators to modify the database query logic. This allows reading checkpoint or store data beyond the intended scope, particularly in multi-tenant deployments where filters enforce tenant isolation. The attack requires authentication and network access to the API, but no user interaction. Patches are available: upgrade langgraph-checkpoint-mongodb to ≥0.3.0 and langgraph-store-mongodb to ≥0.4.0. As a mitigation, strip or escape "$" characters from user-supplied filter inputs before passing them to vulnerable methods.

Affected products

  • LangChain langgraph-checkpoint-mongodb < 0.3.0
  • LangChain langgraph-store-mongodb < 0.4.0

Identifiers

  • CVE-2026-55253
  • GHSA-533j-2v4q-mw5h
  • PYSEC-2026-3859
  • PYSEC-2026-3860

Timeline

  • 2026-06-05: disclosed: Initial advisory published
  • 2026-06-11: patched: Patches released: langgraph-checkpoint-mongodb 0.3.0 and langgraph-store-mongodb 0.4.0
  • 2026-08-20: advisory: GitHub Advisory GHSA-533j-2v4q-mw5h published; CVE-2026-55253 assigned

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.