EUVD-2026-16913Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A07:2025 - Authentication Failures
ASVS 5.0
V6 - Authentication Credential externalization
MITRE ATT&CK
2 techniques across 2 tactics: Persistence, Credential Access The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Impact | This vulnerability presents a mismatch between its low CVSS score (3.3) and real-world risk. CVSS factors (AV:L local-only, PR:L requires low privilege, C:L confidentiality impact, no integrity/availability impact) correctly classify it as low-severity in isolation. However, EPSS exploitation probability is not provided, and several aggravating factors elevate practical risk: publicly available exploit code (POC:YES) lowers exploitation barriers, local access is a common post-compromise step in multi-stage attacks (making this a privilege escalation vector), vendor non-responsiveness eliminates any timeline for official remediation, and hard-coded credentials often chain into lateral movement or cloud account compromise. This is best classified as a low-urgency but high-context-dependent issue-critical for environments with multi-tenant or untrusted-user access to the same system, low priority for isolated development installations. |
| Exploit ScenarioDerived | An attacker with local system access (e.g., a compromised container, shared CI/CD runner, or insider threat) executes the publicly available exploit from https://gist.github.com/YLChen-007/3bf37486022d4c57caec3a35cd79ac92, which reads or extracts the LITELLM_MASTER_KEY from the config.py file or process memory. The attacker then uses this stolen key to authenticate to LiteLLM APIs as the OpenUI application, enabling unauthorized API calls, model inference, or potential lateral movement into downstream language model services or cloud accounts. This attack requires no elevated privileges (PR:L), no user interaction (UI:N), and no additional attack complexity (AC:L), making it trivial once local access is established. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Mar 28, 2026
Wandb OpenUI up to version 1.0 contains hard-coded credentials exposure in backend/openui/config.py where the LITELLM_MASTER_KEY argument is improperly handled, allowing local authenticated users with low privileges to read sensitive authentication material. The vulnerability has a low CVSS score (3.3) due to local-only attack vector and low impact scope, but publicly available exploit code exists and vendor contact has been unsuccessful, increasing practical risk for deployed instances.
RemediationAI
No vendor-released patch identified at time of analysis. Immediate workarounds include: (1) restrict local system access to OpenUI deployment environments using OS-level permissions and user isolation, (2) avoid storing LITELLM_MASTER_KEY in config.py or other checked-in files-use environment variables or secure secrets management systems (e.g., AWS Secrets Manager, HashiCorp Vault) instead, (3) implement file integrity monitoring on backend/openui/config.py to detect unauthorized access, and (4) rotate all LiteLLM credentials if local compromise is suspected. Monitor the wandb GitHub repository and VulDB references (https://vuldb.com/?id.353880) for eventual vendor response or community-maintained patches.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
The vulnerability stems from improper secrets management in OpenUI's configuration layer (CWE-798: Use of Hard-coded Credentials). The LITELLM_MASTER_KEY parameter, which should be a sensitive authentication token for LiteLLM API interactions, is being stored or handled in a way that exposes it to local file system access or process memory inspection. OpenUI is a wandb-integrated tool for UI generation and backend configuration, making the config.py file a critical attack surface. The flaw exists across wandb OpenUI versions up to and including 1.0, with no explicit version cutoff indicating remediation.
Know when Wandb 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
CVSS:4.0/AV:L/AC:L/AT:N/PR:L/UI:N/VC:L/VI:N/VA:N/SC:N/SI:N/SA:N/E:P
Primary rating from NVD · only source for this CVE.
CVSS VectorNVD
Lifecycle Timeline
2DescriptionCVE.org
A vulnerability has been found in wandb OpenUI up to 0.0.0.0/1.0. This impacts an unknown function of the file backend/openui/config.py. The manipulation of the argument LITELLM_MASTER_KEY leads to hard-coded credentials. An attack has to be approached locally. The exploit has been disclosed to the public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.
Executive ViewAI
CVE-2026-4993 is a low-severity vulnerability in wandb OpenUI (CVSS 1.9). The limited severity suggests constrained impact, typically requiring specific conditions or local access for exploitation. Apply vendor updates when available as part of routine maintenance.
Apply vendor patch during routine maintenance. Standard security controls provide adequate interim protection.
Threat Intelligence Overview
NVD
CVE.org
EPSS
Vulnrichment
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.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Authentication bypass attacks exploit flaws in the verification mechanisms that control access to systems and applications.
Related Vulnerabilities in wandb OpenUI Multiple Vulnerabilities
View all 3Path traversal in the Weights & Biases (wandb) Python client before 0.29.0 lets a malicious or compromised W&B backend p
HTML injection in wandb OpenUI up to version 1.0 allows remote unauthenticated attackers to inject arbitrary HTML via ma
Artifact Integrity Validation in wandb 0.25.2.dev1 uses MD5 (a cryptographically broken hash algorithm) within ArtifactM
Exploit-DB