Technology
4 synchronized stories carry this tag.
vLLM fixed a critical flaw affecting versions from 0.8.3 through 0.14.0 where a malicious remote video URL could execute code before API-key authentication protected the request path.
Why it matters: Public multimodal inference servers often fetch remote media, so an input URL could cross the network boundary and compromise the model host before ordinary API authorization.
React disclosed CVE-2025-55182, a maximum-severity flaw allowing unauthenticated remote code execution through vulnerable Server Components payload handling. Affected React 19 server-dom packages could expose applications even without user-authored Server Functions; CISA later confirmed exploitation.
Why it matters: The vulnerable packages sit underneath popular web frameworks used for AI applications and administrative surfaces. Remote compromise could expose prompts, credentials, model access, customer data, source code, and deployment infrastructure.
NVIDIA's August bulletin disclosed three critical network-reachable Triton Inference Server vulnerabilities involving remote code execution, reverse-shell, and stack-overflow paths, with release 25.07 providing the complete fixes.
Why it matters: Triton commonly sits on the network boundary of GPU inference systems, so a server flaw can expose model hosts, credentials, shared memory, and adjacent services.
CISA added the Langflow code-injection vulnerability to its Known Exploited Vulnerabilities catalog after exploitation was observed; Langflow versions before 1.3.0 were affected.
Why it matters: An internet-reachable AI workflow builder could give an attacker code execution without authentication, turning a visual development surface into an infrastructure foothold.