---
id: CVE-2026-103270
title: >-
  LightLLM through 1.2.0 mounts reinforcement learning control routes on the
  public HTTP API without authentication checks
summary: >-
  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_weight…
severity: high
cvss: 7.5
cvssVector: 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H'
cwe:
  - CWE-306
vendor: ModelTC
product: LightLLM
affected:
  - LightLLM <= 1.2.0
published: '2026-09-30'
updated: '2026-09-30'
sourceUpdated: '2026-09-30T16:17:08.847'
source: NVD
sourceUrl: 'https://nvd.nist.gov/vuln/detail/CVE-2026-103270'
references:
  - url: 'https://github.com/ModelTC/LightLLM'
    label: disclosure@vulncheck.com
  - url: 'https://github.com/ModelTC/LightLLM/issues/1609'
    label: disclosure@vulncheck.com
  - url: >-
      https://github.com/ModelTC/lightllm/blob/v1.2.0/lightllm/server/api_http.py#L505-L507
    label: disclosure@vulncheck.com
  - url: >-
      https://github.com/ModelTC/lightllm/blob/v1.2.0/lightllm/server/api_http_rl.py#L51-L139
    label: disclosure@vulncheck.com
  - url: >-
      https://www.vulncheck.com/advisories/lightllm-through-1.2.0-missing-authentication-on-rl-control-routes
    label: disclosure@vulncheck.com
  - url: 'https://github.com/ModelTC/LightLLM/issues/1609'
    label: 134c704f-9b21-4f2e-91b3-4a467353bcc0
tags:
  - nvd
  - cve.org
  - exploit-available
exploitAvailable: true
ssvc:
  exploitation: poc
  automatable: 'yes'
  technicalImpact: partial
  timestamp: '2026-09-30T15:58:00.673012Z'
ingestedAt: '2026-09-30T15:07:05.378Z'
---

## Overview

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.

## Remediation

Refer to the linked advisories for vendor-supplied fixes and affected version ranges.