CVE-2026-40217

CVE Enrichment

CVSS: 8.8 HIGH
EPSS: 15.53%
EPSS Percentile: 96.83%
CISA KEV: Not listed

Description

LiteLLM through 2026-04-08 allows remote attackers to execute arbitrary code via bytecode rewriting at the /guardrails/test_custom_code URI.

Technical Details

CVSS Version: 3.1
CNA / NVD Source: [email protected]
Published: 10-Apr-2026 21:16
Last Modified: 27-Oct-2026 04:10
EPSS Date: 2026-10-30
CWE: CWE-420, CWE-94

CVSS Vector


            CVSS:3.1                            AV:N /
                            AC:L /
                            PR:L /
                            UI:N /
                            S:U /
                            C:H /
                            I:H /
                            A:H

Affected Products / Versions

litellm / litellm Source: NVD CPE · affected
Version / affected range: * · through (inclusive) 2026-04-08
cpe:2.3:a:litellm:litellm:*:*:*:*:*:*:*:*

References

Last enrichment sync: 31-Oct-2026 08:25
Sources: NVD · FIRST EPSS · CISA KEV
Latest successful provider update; providers refresh independently.

Reported Context (3)

The researchers have also observed threat actors exploiting CVE-2026-40217, a vulnerability in LiteLLM that allows running arbitrary code on the server by leveraging the custom-code guardrail testing endpoint. LLMjacking Attacks Drive Cloud AI Charges Above $1 Million, Google Says
The actor exploited a user ID rebind vulnerability in the /key/regenerate endpoint (CVE-2026-40217) to elevate standard Virtual Keys into full PROXY_ADMIN master administrative credentials, establishing deep, persistent Mandiant Details LLMJacking Attacks Monetizing Compromised Cloud LLM Access
#LiteLLM Three LiteLLM Vulnerabilities Could Let Low-Privilege Users Take Over AI Gateways

CVE (3)

Threat Actors (1)

MITRE ATT&CK (20)

Vendors (4)

Products (6)

Tools (6)

Industries (4)

Note: Related entities, including threat actors, malware, CVEs, MITRE ATT&CK techniques, vendors, products, tools, countries, and industries, are shown when they appear in the same reporting. Their presence does not necessarily mean they were targeted, compromised, vulnerable, responsible for the activity, or directly involved in the incident.