PT-2026-7475 · Langchain · Langchain

Finder16 · Published 2026-02-10 · Updated 2026-02-16 · CVE-2026-26013

Positive Technologies (AI Summary)

CVSS v3.1

3.7

Low

Base vector

AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:L

Name of the Vulnerable Software and Affected Versions LangChain versions prior to 1.2.11

Description LangChain is a framework used for building applications powered by large language models (LLMs). A flaw exists in the ChatOpenAI.get num tokens from messages() method where it retrieves image URLs without proper validation when calculating token counts for models with vision capabilities. This can be exploited to initiate Server-Side Request Forgery (SSRF) attacks. Attackers can achieve this by supplying malicious image URLs as part of user-provided input. The vulnerable function is ChatOpenAI.get num tokens from messages(). The vulnerable parameter is image url.

Recommendations Update to LangChain version 1.2.11 or later.

Exploit

Fix

SSRF

Weakness Enumeration

CWE-918

Related Identifiers

CVE-2026-26013

GHSA-2G6R-C272-W58R

Affected Products

Langchain

References · 13

• 🔥 https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/26xxx/CVE-2026-26013.json · Exploit
• https://github.com/langchain-ai/langchain/commit/2b4b1dc29a833d4053deba4c2b77a3848c834565 · Patch
• https://github.com/langchain-ai/langchain/security/advisories/GHSA-2g6r-c272-w58r · Vendor Advisory
• https://nvd.nist.gov/vuln/detail/CVE-2026-26013 · Security Note
• https://osv.dev/vulnerability/GHSA-2g6r-c272-w58r · Vendor Advisory
• https://osv.dev/vulnerability/CVE-2026-26013 · Vendor Advisory
• https://github.com/langchain-ai/langchain · Note
• https://github.com/langchain-ai/langchain/releases/tag/langchain-core%3D%3D1.2.11 · Note
• https://twitter.com/VulmonFeeds/status/2021421627607503248 · Twitter Post
• https://t.me/CVEtracker/44360 · Telegram Post
• https://twitter.com/CVEnew/status/2021365281814106374 · Twitter Post
• https://twitter.com/CisoRaging77913/status/2023266252215472212 · Twitter Post
• https://twitter.com/_cvereports/status/2021595099264700877 · Twitter Post