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.
{
"cna_assigner": "VulnCheck",
"cwe_ids": [
"CWE-306"
],
"osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/103xxx/CVE-2026-103270.json"
}{
"extracted_events": [
{
"introduced": "0"
},
{
"last_affected": "1.2.0"
},
{
"fixed": "1.2.0"
}
],
"source": [
"AFFECTED_FIELD",
"DESCRIPTION"
]
}