LiteLLM through 2026-04-08 allows remote attackers to execute arbitrary code via bytecode rewriting at the /guardrails/test_custom_code URI.
| Vendor | Product | Versions |
|---|---|---|
| litellm | litellm | bb0639701796218a3447160e55c0f1097446e4e6085df7dfd39f476d4143743f |
Downstream vendors/products affected by this vulnerability
| Vendor | Product | Source | Confidence |
|---|---|---|---|
| open source | litellm | cert_advisory | 90% |
| red hat | ansible_automation_platform | cve_cpe | 95% |
| red hat | openshift_ai | cve_cpe | 95% |
| red hat | openshift_lightspeed | cve_cpe | 95% |
Loading…
Added MITRE ATT&CK techniques (T1190 for exploit, T1078/T1078.004 for credential access, T1087 for account discovery, T1555 for credentials from password managers, T1552 for credential exposure) based on attack methodology described.
Added MITRE ATT&CK techniques (T1190 for exploitation, T1528 for stolen credentials, T1556 for modify cloud compute infrastructure) and new tags documenting LLMjacking attack campaigns and threat actor UNC6993 exploiting this CVE in the wild.
Updated description to include that the vulnerability allows execution of arbitrary code with administrator privileges and changed severity from HIGH to CRITICAL.
Updated severity to CRITICAL and marked the vulnerability as actively exploited.
Initial creation