LangGraph patches three flaws, stops RCE chain
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Check Point disclosed three now‑patched security flaws in LangGraph, including a critical chain that can lead to remote code execution.
- CVE-2025-67644 (CVSS 7.3) is an SQL injection in LangGraph's SQLite checkpoint implementation, affecting langgraph‑checkpoint‑sqliteite versions before 3...1.
- CVE-2026-28277 (CVSS 6.8) is an unsafe msgpack deserialization vulnerability in LangGraph, affecting langgraph versions before 1.0.10.
- CVE-2026-27022 (CVSS 6.5) is a RediSearch query injection in @langchain/langgraph‑checkpoint‑redis, affecting @langchain/langgraph‑checkpoint‑redis versions before 1.0.1.
- Yarden Porat discovered the flaws and noted that CVE‑2025‑67644 and CVE‑2026‑28277 can be chained via the get_state_history() endpoint to achieve remote code execution on self‑hosted deployments.
- LangChain's LangSmith managed platform is not affected, while self‑hosted servers using SQLite or Redis checkpointers remain vulnerable unless patched.
- LangGraph maintainers advise users to apply the fixes, enable authentication, enforce network segmentation, and follow the principle of least privilege to mitigate risk.
Why it matters: Self‑hosted LangGraph operators must promptly apply the patches, otherwise attackers can exploit the SQL‑injection and deserialization chain to run arbitrary code, exposing runtime secrets and compromising linked systems. The managed LangSmith service remains unaffected, shifting the security burden to on‑premise users.
