CVE-2026-34070: LangChain is a framework for building agents and LLM-powered applications. Prior to version 1.2.22, …
- Occurred
- Mar 31, 2026
- Published
- Mar 31, 2026
- Source
- NVD CVE
Summary
LangChain is a framework for building agents and LLM-powered applications. Prior to version 1.2.22, multiple functions in langchain_core.prompts.loading read files from paths embedded in deserialized config dicts without validating against directory traversal or absolute path injection. When an application passes user-influenced prompt configurations to load_prompt() or load_prompt_from_config(), an attacker can read arbitrary files on the host filesystem, constrained only by file-extension checks (.txt for templates, .json/.yaml for examples). This issue has been patched in version 1.2.22.
Tags
- CVE
- CVE-2026-34070
- cvss:HIGH
Classification
Rationale: Vulnerability allows unauthorized file access via prompt configuration injection, enabling data exfiltration.
Classifier: Qwen/Qwen3.5-35B-A3B-FP8