Is CVE-2026-105699 exploited in the wild?

No. Junglewise Threat Intelligence has no report of CVE-2026-105699 being exploited in the wild.

How severe is CVE-2026-105699?

High severity, with a CVSS score of 4, published 5 October 2026.

Which products and versions are affected?

Langflow-Ai Langflow 1.6.8 to 1.9.0; langflow (PyPI) >= 1.6.8, <= 1.9.0, ECOSYSTEM:introduced=1.6.8,fixed=1.9.1.

Is there a fix for CVE-2026-105699?

Yes. langflow (PyPI): upgrade to 1.9.1 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

How is CVE-2026-105699 attacked?

It is attacked over the network.

Am I affected by CVE-2026-105699?

Check whether you run Langflow-Ai Langflow or langflow (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.

CVE-2026-105699: Langflow path traversal and cross-user disclosure in MCP resources

Q: Is there a fix for CVE-2026-105699?

Yes. langflow (PyPI): upgrade to 1.9.1 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

Q: How is CVE-2026-105699 attacked?

It is attacked over the network.

Q: Am I affected by CVE-2026-105699?

Check whether you run Langflow-Ai Langflow or langflow (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.

CVE-2026-105699 · Severity: high · CVSS 4 · Published 2026-10-05

Langflow is a platform for building AI-powered workflows and agents. Authenticated users could read arbitrary files belonging to other users or projects through the Model Context Protocol (MCP) resources endpoint by bypassing authorization checks. This could expose uploaded documents, prompts, structured data, and other private workflow artifacts, though without the ability to modify them.

Technologies: Langflow-Ai Langflow, langflow (PyPI). Vendors: Langflow-Ai, PyPI.

Quick answers

Is CVE-2026-105699 exploited in the wild?: No. Junglewise Threat Intelligence has no report of CVE-2026-105699 being exploited in the wild.
How severe is CVE-2026-105699?: High severity, with a CVSS score of 4, published 5 October 2026.
Which products and versions are affected?: Langflow-Ai Langflow 1.6.8 to 1.9.0; langflow (PyPI) >= 1.6.8, <= 1.9.0, ECOSYSTEM:introduced=1.6.8,fixed=1.9.1.
Is there a fix for CVE-2026-105699?: Yes. langflow (PyPI): upgrade to 1.9.1 or later. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.
How is CVE-2026-105699 attacked?: It is attacked over the network.
Am I affected by CVE-2026-105699?: Check whether you run Langflow-Ai Langflow or langflow (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

langflow (PyPI): upgrade to 1.9.1 or later
```
pip install "langflow>=1.9.1"
```

Technical details

The vulnerability combines path traversal with broken access control: the MCP resources/read handler accepted user-supplied filenames and forwarded them to storage_service.get_file without validating ownership or project membership. Path-containment checks (rejecting ../, /, \) were missing from read operations, and list operations returned flows from all users regardless of authentication context. An authenticated attacker with any project-scoped MCP access could request another user's files, and global list endpoints disclosed identifiers making targeting easier.

Affected products

Langflow AI Langflow 1.6.8 to 1.9.0

Identifiers

CVE-2026-105699
GHSA-4hmc-cfm3-w43c
PYSEC-2026-4230

Timeline

2026-10-05: disclosed
2026-04-21: patched

References

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