IP Spoofing Vulnerability in Langflow Tool by Langflow
CVE-2026-105741

7.1HIGH

Key Information:

Vendor

Langflow-ai

Status
Langflow
Vendor
CVE Published:
5 October 2026

What is CVE-2026-105741?

Langflow, a tool designed for developing AI-powered agents and workflows, has an IP spoofing vulnerability in the Model Context Protocol (MCP) installation endpoint. Versions 1.5.0 to 1.10.3 are affected, allowing authenticated remote attackers to exploit this weak point by spoofing an X-Forwarded-For header. This manipulation can enable attackers to bypass the designated 'local-only' access restriction, permitting unauthorized writing or overwriting of the MCP client configuration file on the server. This vulnerability was addressed in the 1.10.3 release.

Affected Version(s)

langflow >= 1.5.0, < 1.10.3

References

https://github.com/langflow-ai/langflow/security/advisori...
x_refsource_CONFIRM
https://github.com/langflow-ai/langflow/pull/13915
x_refsource_MISC
https://github.com/langflow-ai/langflow/commit/1f39a4b9d6...
x_refsource_MISC

CVSS V3.1

Score:
7.1
Severity:
HIGH
Confidentiality:
None
Integrity:
High
Availability:
None
Attack Vector:
Network
Attack Complexity:
Low
Privileges Required:
Low
User Interaction:
None
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.