LangChain JS serialization injection vulnerability CVE-2025-68665

# LangChain JS serialization injection vulnerability CVE-2025-68665

https://gcdradar.vercel.app/incidents/langchain-js-serialization-injection-vulnerability-cve-2025 · AIR-2025-0389 · closed 2025-12-23

- Lane: Patch now
- Severity: 2 (Disclosed vulnerability)
- Fix status: Fix status unknown
- First reported: 2025-12-23
- AI role: Attack on AI
- Products: LangChain JS
- CVE ids: CVE-2025-68665
- CVSS: 8.6

## Summary

A serialization injection vulnerability in LangChain JS fails to escape objects containing lc keys during serialization. User-controlled data containing this structure is improperly treated as a legitimate LangChain object during deserialization.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2025-12-23: CVE-2025-68665: LangChain is a framework for building LLM-powered applications. (https://nvd.nist.gov/vuln/detail/CVE-2025-68665)
- 2025-12-23: CVE-2025-68665: LangChain serialization injection vulnerability enables secret extraction (https://github.com/advisories/GHSA-r399-636x-v7f6)

## Sources

- 2025-12-23 GitHub Advisory Database: CVE-2025-68665: LangChain serialization injection vulnerability enables secret extraction https://github.com/advisories/GHSA-r399-636x-v7f6
- 2025-12-23 NVD: CVE-2025-68665: LangChain is a framework for building LLM-powered applications. https://nvd.nist.gov/vuln/detail/CVE-2025-68665

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.