[CVE-2026-103270] LightLLM through 1.2.0 mounts reinforcement learning control routes on the public HTTP API without a…

Descripción técnica

LightLLM through 1.2.0 mounts reinforcement learning control routes on the public HTTP API without authentication checks. Unauthenticated attackers can call endpoints like /pause_generation, /abort_request, /flush_cache, and /init_weights_update_group to disrupt inference operations and wedge workers on deployments started with --enable_rl.

Puntuación CVSS

Score: 7.5/10 — Severidad: HIGH — Estado NIST: Deferred

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Debilidades (CWE)

CWE-306

Fuente oficial

Publicado en NIST NVD.