Skip to main content

LiteLLM EUVDEUVD-2026-42361

| CVE-2026-59819 LOW
External Control of File Name or Path (CWE-73)
2026-07-08 GitHub_M GHSA-4g5m-c9r5-49xf
2.1
CVSS 4.0 · Vendor: GitHub_M
View

Attack ChainAIDerived

Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain

  1. Access
    Obtain proxy admin credentials
  2. Delivery
    Authenticate to LiteLLM management API
  3. Exploit
    Craft litellm_params with oidc/file/ path reference
  4. Execution
    POST to /health/test_connection endpoint
  5. Persist
    Server resolves path and reads local file
  6. Impact
    Exfiltrate returned file contents

Vulnerability Assessment

Weakness
CWE
OWASP Top 10:2025
A06:2025 - Insecure Design
ASVS 5.0
V5 - File Handling File name/path control
Exploitation Conditions

Exploitation requires the attacker to hold a proxy administrator role or equivalent privileged caller permission that grants access to the /health/test_connection endpoint - this is explicitly stated in the vulnerability description as the required privilege level. The CVSS 4.0 AT:P metric further indicates a specific attack requirement beyond normal configuration: the caller must have permission specifically to test model connections, which may be a separately grantable capability in LiteLLM's access model. Limiting factors include: (1) PR:H - standard authenticated user accounts without admin rights cannot trigger this; (2) AT:P - even among admins, test-connection permission may not be universally granted; (3) impact is read-only (VC:L), with no integrity or availability consequence. Exploitation is not automatable at scale and requires an attacker who has already achieved administrative access to the LiteLLM proxy.

Impact

Real-world risk is low, well-aligned with the CVSS 4.0 score of 2.1. The CVSS 4.0 vector (AV:N/AC:L/AT:P/PR:H/UI:N/VC:L/VI:N/VA:N) highlights two critical limiters: PR:H requires the attacker to already be an authenticated proxy administrator or another caller explicitly granted test-connection permission, and AT:P signals an additional attack requirement beyond standard configuration (the caller must have access to the specific /health endpoint and permission to invoke it). The VC:L rating reflects that file contents may be limited in sensitivity depending on filesystem permissions and deployment layout. No EPSS data was provided in the intelligence feed; however, given PR:H and the specialized attacker profile required, opportunistic mass exploitation is implausible. This vulnerability is absent from the CISA KEV catalog and no public exploit code has been identified. The primary realistic risk is insider threat or a compromised administrator account pivoting to credential file exfiltration - for example, reading service account tokens or API key files stored on the proxy host.

Exploit ScenarioDerived

An authenticated proxy administrator logs into the LiteLLM management interface and invokes /health/test_connection with a crafted litellm_params payload containing an oidc/file//etc/passwd or oidc/file//path/to/api_keys reference. The server resolves the path against the local filesystem and returns the file contents within the connection test response, allowing the attacker to harvest credentials or sensitive configuration files stored on the proxy host. No public exploit code has been identified; exploitation requires valid administrative credentials.

Monitor Track and reassess. Why: LOW severity (CVSS 0.1-3.9).
Patch available 0.3% exploitation probability (EPSS)

Get an e-mail when this CVE is exploited, patched or re-scored. Free account.

Alert me

AnalysisAIv1 – Jul 08, 2026

Local filesystem read in LiteLLM's AI Gateway proxy prior to version 1.83.10-stable allows a proxy administrator to exfiltrate files from the server by supplying crafted oidc/file/ path references through the /health/test_connection endpoint. The vulnerability stems from unsanitized resolution of request-supplied environment and OIDC file references within litellm_params, giving a high-privileged caller a direct path to arbitrary file read. No public exploit code or active exploitation has been identified; the low CVSS 4.0 score of 2.1 reflects the high privilege bar and limited confidentiality impact.

RemediationAI

Upgrade to LiteLLM version 1.83.10-stable or later, which contains the fix as described in GitHub Security Advisory GHSA-4g5m-c9r5-49xf and merged via pull request https://github.com/BerriAI/litellm/pull/25592; the release is available at https://github.com/BerriAI/litellm/releases/tag/v1.83.10-stable. If immediate upgrade is not feasible, restrict access to the /health/test_connection endpoint at the network perimeter or API gateway layer to only trusted administrator IP ranges, reducing the pool of callers who could supply malicious litellm_params. Additionally, review and tighten proxy administrator role membership, applying the principle of least privilege so that only operationally required accounts hold test-connection permission. Note that network-layer blocking of /health/test_connection will prevent legitimate connectivity diagnostics - assess operational impact before applying this workaround.

Affected VersionsENISA

LiteLLM < 1.83.10-stable

Affected Packages

pypi litellm fix: 1.83.10-stable, 1.83.10

Source: OSV (Open Source Vulnerabilities)

Affects your stack

Affects 1 package in pypi

Checking this CVE against your own stacks and servers is part of the Basic plan and up.

Technical ContextAI · show details

LiteLLM (CPE: cpe:2.3:a:berriai:litellm:*:*:*:*:*:*:*:*) is an open-source Python-based AI Gateway that proxies requests to multiple LLM provider APIs using OpenAI-compatible formatting. The affected component is its /health/test_connection endpoint, which accepts model configuration parameters (litellm_params) and uses them to validate connectivity. CWE-73 (Improper Neutralization of External Input Used by a Downstream Component, also known as External Control of File Name or Path) describes the root cause: user-supplied strings in litellm_params are passed directly to a file resolution routine that supports both environment variable substitution and OIDC credential file references via the oidc/file/ URI scheme. The server resolves these references against the local filesystem without restricting the path, effectively turning the connection test into a server-side file read primitive.

Know when Litellm gets its next vulnerability

  • Watch products and get an email when something new is published
  • Check this CVE against your own stack and servers
  • Exploit and PoC references, threat intelligence from 22 sources

No credit card · 7-day full trial

Severity by source

Vendor (GitHub_M) PRIMARY 2.1 LOW
CVSS:4.0/AV:N/AC:L/AT:P/PR:H/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N
vuln.today AI 2.7 LOW

Network-reachable endpoint but requires high privileges (proxy admin); confidentiality impact is low as only partial file read applies with no integrity or availability consequence.

3.1 AV:N/AC:L/PR:H/UI:N/S:U/C:L/I:N/A:N
4.0 AV:N/AC:L/AT:P/PR:H/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N

Primary rating from Vendor (GitHub_M).

CVSS VectorVendor: GitHub_M

Attack Vector
Network
Attack Complexity
Low
Attack Requirements
Present
Privileges Required
High
User Interaction
None
Vulnerable system: Confidentiality
Low
Vulnerable system: Integrity
None
Vulnerable system: Availability
None
Subsequent system: Confidentiality
None
Subsequent system: Integrity
None
Subsequent system: Availability
None

Lifecycle Timeline

1
Patch available
Jul 08, 2026 - 21:04 EUVD

DescriptionCVE.org

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. Prior to 1.83.10-stable, LiteLLM's /health/test_connection endpoint resolved request-supplied environment and OIDC file references in litellm_params, allowing a proxy administrator or another privileged caller with permission to test model connections to read files from the local filesystem via an oidc/file/ reference. This issue is fixed in version 1.83.10-stable.

Blast Radius

ecosystem impact
† from your stack dependencies † transitive graph · vuln.today resolves 4-path depth
  • 15 pypi packages depend on litellm (14 direct, 1 indirect)

Ecosystem-wide dependent count for version 1.83.10.

Threat Intelligence Overview

EPSS Score
0.3%
Top 80% exploitable
Prevalence
Medium
popular AI gateway in LLM-heavy enterprise and startup deployments
Threat Intel
1
1 source
Data Quality
7 sources
Confidence: High
EUVD NVD CVE.org EPSS Vulnrichment GHSA GitHub VulDB

How this decision was reached

The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.

SignalValue
Severity LOW
CISA KEV No
Public POC No
EPSS 0.3%
Patch available Yes
Active threat No
Deciding rule
LOW severity (CVSS 0.1-3.9)
What every action level means
Emergency Immediate action required. Criteria: CRITICAL + in CISA KEV (actively exploited) without patch, OR CRITICAL + EPSS > 30% without patch, OR CRITICAL + active threat signals without patch.
Act Now Prioritize within 24-48h. Criteria: in KEV with patch available, OR CRITICAL + public POC, OR HIGH + EPSS > 10%, OR CRITICAL + active threat, OR any CRITICAL severity.
This Week Schedule remediation this week. Criteria: HIGH severity (CVSS 7.0-8.9). Patch within normal change window.
This Month Plan within 30 days. Criteria: MEDIUM severity (CVSS 4.0-6.9), lower exploitation probability. Include in regular patch cycle.
Monitor Track and reassess. Criteria: LOW severity (CVSS 0.1-3.9), minimal exploitation risk. No immediate action needed.

SSVC AssessmentENISA

SSVC Decision· derived
Defer
Exploitation
None
System Exposure· derived
Open
Automatable
No
Human Impact· derived
Low
Technical Impact
Partial

Priority Score

10
Low Medium High Critical
KEV: 0
EPSS: 0
CVSS: +10
PoC: 0
Weaponized: 0
Threat intel: 0
Gov advisory: 0

Attack Technique

Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security.

CVE-2026-42271 HIGH POC
8.7 May 08

Remote command execution in LiteLLM proxy server versions 1.74.2 through 1.83.6 allows any authenticated user to execute

CVE-2026-42208 CRITICAL POC
9.3 May 08

SQL injection in LiteLLM proxy server versions 1.81.16 through 1.83.6 allows unauthenticated remote attackers to read an

CVE-2026-59822 HIGH POC
8.8 Jul 08

Authentication bypass in LiteLLM (BerriAI) proxy/AI Gateway before 1.84.0 lets a remote unauthenticated attacker reach M

CVE-2024-2952 CRITICAL POC
9.8 Apr 10

BerriAI/litellm is vulnerable to Server-Side Template Injection (SSTI) via the `/completions` endpoint. Rated critical s

CVE-2026-40217 HIGH POC
8.8 Apr 10

Remote code execution in BerriAI LiteLLM (all versions through 2026-04-08) enables authenticated attackers to execute ar

CVE-2024-4888 HIGH POC
8.1 Jun 06

BerriAI's litellm, in its latest version, is vulnerable to arbitrary file deletion due to improper input validation on t

CVE-2026-93355 HIGH POC
7.6 Sep 28

Authentication bypass in the LiteLLM proxy lets any party who already holds a valid JWT issued by the configured externa

CVE-2025-0330 HIGH POC
7.5 Mar 20

In berriai/litellm version v1.52.1, an issue in proxy_server.py causes the leakage of Langfuse API keys when an error oc

CVE-2024-9606 HIGH POC
7.5 Mar 20

In berriai/litellm before version 1.44.12, the `litellm/litellm_core_utils/litellm_logging.py` file contains a vulnerabi

CVE-2024-6587 HIGH POC
7.5 Sep 13

A Server-Side Request Forgery (SSRF) vulnerability exists in berriai/litellm version 1.38.10. Rated high severity (CVSS

CVE-2024-5225 HIGH POC
7.2 Jun 06

An SQL Injection vulnerability exists in the berriai/litellm repository, specifically within the `/global/spend/logs` en

CVE-2024-4889 HIGH POC
7.2 Jun 06

A code injection vulnerability exists in the berriai/litellm application, version 1.34.6, due to the use of unvalidated

This site uses cookies essential for authentication and security. No tracking or analytics cookies are used. Privacy Policy