Vulnerabilities in Lang Chain Gen AI
Summary: Palo Alto Networks researchers identified two vulnerabilities in LangChain, the open source generative AI framework with over 81,000 GitHub stars and more than one million builders: CVE-2023-46229, an SSRF in versions before 0.0.317 where a crafted sitemap could pull sensitive intranet data, and CVE-2023-44467, a critical prompt injection flaw in LangChain Experimental before 0.0.306 that let attackers run arbitrary code through PALChain. LangChain patched both, adding a domain allowlist for the sitemap loader and expanding the code blocklist in pull request langchain-ai/langchain#11233.
Key Takeaway: The PALChain blocklist covered only four functions - system, exec, execfile and eval - and researchers bypassed it with Python's built-in __import__() to load subprocess and achieve remote code execution, a reminder that a framework sitting in front of your LLM can turn a crafted prompt into a shell command.
Read the original at Unit 42 Palo Alto Networks ↗
- Topics
- Vulnerabilities & Patching
- Mentions
- CVE-2023-44467CVE-2023-46229
- Related
- Attackers Exploit UnrealIRCd Flaw That Enables Server-Level Command Injection Cyber Risks · Nov 14, 2026U.S. CISA adds Check Point and UnrealIRCd flaws to its Known Exploited Vulnerabilities catalog Cyber Risks · Nov 14, 2026PostgreSQL Patches 52 Flaws, Including 22-Year-Old Bug Enabling Authenticated RCE Cyber Risks · Nov 13, 2026