CVE Professor
›
Back
CVE-2026-103270

CVE-2026-103270 — LightLLM — CWE-306

HIGH CVSS 7.5 NVD

CVE-2026-103270 is a HIGH severity Missing Authorization & Access Control vulnerability (CWE-306) (CVSS 7.5/10) affecting LightLLM. Technical vector analysis shows that an attacker can exploit this flaw remotely over the network without authentication (PR:N) with zero user interaction (UI:N), causing high-impact service disruption (A:H).

📄 Official NVD / MITRE Original Record Text

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.

Published 30.09.2026
ISO Date 2026-09-30
CVSS Version v3.1
Source NVD
CVSS v3.1 Vector Metrics
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Attack Vector Network AV:N
Attack Complexity Low AC:L
Privileges Req. None PR:N
User Interaction None UI:N
Scope Unchanged S:U
Confidentiality None C:N
Integrity None I:N
Availability High A:H
AI Security Analysis & Remediation
🤖 Generating Deep AI Security Analysis…
Analyzing CVSS vector & CWE weakness structure…

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
CVE-2026-103270
CVSS 7.5/10 · HIGH · 30.09.2026
CWE: CWE-306
Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

🎯 WHAT IS THIS VULNERABILITY?
Authentication Bypass
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_weights_update_group to disrupt inference operations and wedge workers on deployments started with --enable_rl.

⚡ ATTACK CHAIN
1. JWT: change alg to none, remove signature.
2. SQL bypass: ' OR '1'='1'-- in login form.
3. Direct API endpoint access without auth headers.
4. Password reset token prediction or race condition.
5. Credential stuffing with default creds.

🔧 Tools: jwt_tool, Burp Suite, Hydra, sqlmap --forms
🗺️ MITRE ATT&CK: T1078 (Valid Accounts), T1110 (Brute Force)

🌍 REAL WORLD IMPACT
Service can be completely taken down.

🛡️ REMEDIATION
→ Apply the vendor-released patch immediately.
→ Add MFA, strengthen sessions, implement rate limiting.
→ Pre-patch: put service behind VPN or apply IP allowlist.
→ Add WAF rule, create IDS alert for anomalous traffic.

⚠️ URGENCY
⚠️ This week (7 days) — High risk, anonymous remote.

────────────────────────────────────────
[Kural Tabanlı Otomatik Güvenlik Analizi]

Target Reconnaissance Queries (FOFA & Shodan)
CVSS Score
7.5 / 10 · CVSS HIGH
Weakness Types (CWE)
CWE-306
Related CVEs
9.8
CVE-2026-69157
## Summary atomic-server provisions a well-known invite at `{server}/setup` with `write = true`, `target` = t
9.8
CVE-2022-51027
MeterSphere versions 1.13.0 before 1.14.7 and 1.15.0 before 1.16.4 contain a pre-authentication code injection
9.8
CVE-2026-126752
Dagu 1.23.0 before 2.0.0 defaults auth.mode to none, allowing unauthenticated attackers to execute OS commands
9.3
CVE-2026-77297
MyMangaDB is a self-hosted manga collection manager with automatic metadata, MyAnimeList import, Docker suppor
9.3
CVE-2026-77296
MyMangaDB is a self-hosted manga collection manager with automatic metadata, MyAnimeList import, Docker suppor
9.1
CVE-2026-130154
Zyrabit Platform provides sovereign AI infrastructure for enterprise retrieval-augmented generation. Prior to