CVE-2026-44843 LangChain: Unsafe deserialization of attacker-controlled LangChain objects through overly broad `load()` allowlists

LangChain is a framework for building agents and LLM-powered applications. Prior to 0.3.85 and 1.3.3, LangChain contains older runtime code paths that deserialize run inputs, run outputs, or other application-controlled payloads using overly broad object allowlists. These paths may call load with...

CVE-2026-44843

LangChain CVE-2026-44843 affects LangChain-core runtimes prior to 0.3.85 and 1.3.3, which use older code paths with broad object allowlists that can revive trusted LangChain-serializable objects via load()/loads() calls. This may let attacker-controlled serialized constructor dictionaries instant...

Deserialization Of Untrusted Data

Langchain is vulnerable to Deserialization of Untrusted Data. The vulnerability is caused due to improper deserialization of untrusted pickle data in the FAISS.deserializefrombytes function, which allows execution of arbitrary commands via os.system...