✦ CVE-2026-42208 - Critical SQL Injection in LiteLLM

🧭 Overview

CVE-2026-42208 is a critical pre-authentication SQL injection vulnerability affecting the LiteLLM AI gateway.

It allows attackers to exploit backend database queries without authentication, potentially exposing sensitive AI infrastructure data.

📦 Affected Software

• LiteLLM
• Vulnerable versions: 1.81.16 → 1.83.6

🧨 Vulnerability Type

SQL Injection (Pre-authentication)

• No login required 🚫
• Exploitable via HTTP request manipulation
• Targets backend database layer directly

📊 Severity

🔴 Critical

• CVSS Score: ~9.3 – 9.9
• High likelihood of full system compromise if exposed

💥 Potential Impact

If exploited, attackers may:

• 🔐 Extract sensitive API keys (OpenAI, Anthropic, etc.)
• 🗄️ Read or modify database records
• ⚙️ Alter system configuration
• ☁️ Compromise AI gateway infrastructure
• 🚪 Gain foothold for deeper system access

🧬 Demo

📈 Exploitation Status

• 🚨 Actively exploited shortly after disclosure
• ⚡ Rapid weaponization observed in the wild
• 🎯 Targets: exposed LiteLLM deployments

🛠️ Mitigation

✅ Immediate Fix

Upgrade to:

LiteLLM >= 1.83.7 (recommended: 1.83.10 stable)

🔐 Security Recommendations

• Restrict public access to LiteLLM endpoints
• Rotate all exposed API keys immediately
• Monitor logs for unusual Authorization headers
• Enable network-level access controls (firewalls / VPNs)

🧾 Summary

CVE-2026-42208 is a high-impact SQL injection vulnerability in LiteLLM that enables unauthenticated attackers to extract sensitive AI credentials and potentially compromise entire AI infrastructure deployments.