CVE-2023-32786 — In Langchain through 0.0.155, prompt injection allows an attacker to force the service to retrieve data from an arbitrary URL, essentially providing SSRF and potentially injecting content into downstr — CVE Database · The Intelligence Room