CVE-2026-4992: wandb OpenUI XSS Vulnerability

CVE-2026-4992 Overview

A HTML injection vulnerability has been identified in wandb OpenUI up to version 1.0. This security flaw affects the create_share and get_share functions within the file backend/openui/server.py of the HTMLAnnotator Component. By manipulating the ID argument, an attacker can inject malicious HTML content. This attack can be performed remotely over the network and requires some user interaction to be exploited. The exploit has been publicly disclosed and may be actively used. The vendor was contacted about this disclosure but did not respond.

Critical Impact

Remote attackers can inject arbitrary HTML content into the application, potentially leading to content spoofing, phishing attacks, or serving as a stepping stone for more severe cross-site scripting (XSS) attacks.

Affected Products

• wandb OpenUI up to version 1.0
• HTMLAnnotator Component (backend/openui/server.py)

Discovery Timeline

• 2026-03-27 - CVE CVE-2026-4992 published to NVD
• 2026-03-30 - Last updated in NVD database

Technical Details for CVE-2026-4992

Vulnerability Analysis

This vulnerability is classified as CWE-79 (Improper Neutralization of Input During Web Page Generation), commonly known as Cross-Site Scripting or HTML Injection. The flaw exists in the HTMLAnnotator Component of wandb OpenUI, specifically within the create_share and get_share functions located in backend/openui/server.py.

The vulnerability allows remote attackers to inject malicious HTML content by manipulating the ID parameter. When user-supplied input is not properly sanitized before being rendered in the HTML output, attackers can insert arbitrary HTML tags and potentially JavaScript code that will be executed in the context of victims' browsers.

The attack requires network access and some degree of user interaction, meaning a victim must interact with the malicious content for the attack to succeed. This could involve clicking a crafted link or viewing a page containing injected content.

Root Cause

The root cause of this vulnerability is insufficient input validation and sanitization of the ID argument in the create_share and get_share functions. The application fails to properly escape or filter user-controlled input before incorporating it into the generated HTML output, allowing attackers to break out of the intended context and inject arbitrary HTML elements.

Attack Vector

The attack vector is network-based, allowing remote exploitation. An attacker can craft a malicious request containing HTML injection payload in the ID parameter. When this manipulated data is processed by the vulnerable functions and rendered in a web page, the injected HTML content is executed or displayed in the victim's browser.

The attack scenario typically involves:

1. Attacker identifies the vulnerable endpoint handling share functionality
2. Attacker crafts a request with malicious HTML content in the ID parameter
3. The application processes the request without proper sanitization
4. When a victim accesses the affected resource, the injected HTML is rendered in their browser

Technical details and proof-of-concept information are available in the GitHub Gist PoC Repository.

Detection Methods for CVE-2026-4992

Indicators of Compromise

• Unusual HTML tags or script elements appearing in share ID parameters or stored share data
• Unexpected HTTP requests to the create_share or get_share endpoints with suspicious payloads
• Log entries showing ID parameters containing HTML special characters such as <, >, ", or '
• User reports of unexpected content or behavior when viewing shared resources

Detection Strategies

• Implement web application firewall (WAF) rules to detect and block HTML injection patterns in request parameters
• Monitor application logs for requests containing HTML special characters in the ID parameter
• Deploy SentinelOne Singularity platform to detect suspicious web application behavior and potential injection attempts
• Conduct regular security scans of the OpenUI application using DAST (Dynamic Application Security Testing) tools

Monitoring Recommendations

• Enable detailed logging for the backend/openui/server.py endpoints, specifically the share-related functions
• Set up alerts for anomalous patterns in ID parameter values
• Monitor for increased error rates or unusual response sizes from the HTMLAnnotator Component
• Implement real-time security monitoring with SentinelOne to detect exploitation attempts

How to Mitigate CVE-2026-4992

Immediate Actions Required

• Implement strict input validation on the ID parameter to allow only expected characters (alphanumeric values)
• Apply HTML entity encoding to all user-supplied input before rendering in HTML output
• Consider implementing Content Security Policy (CSP) headers to mitigate the impact of potential injection attacks
• If possible, disable or restrict access to the affected share functionality until a patch is available

Patch Information

At the time of publication, no official patch has been released by the vendor. The vendor was contacted about this disclosure but did not respond. Organizations using wandb OpenUI up to version 1.0 should monitor for vendor updates and apply patches as soon as they become available.

For additional technical information, refer to:

• VulDB Entry #353879
• VulDB CTI ID #353879

Workarounds

• Implement a reverse proxy or WAF in front of the application to filter malicious input
• Manually patch the create_share and get_share functions to sanitize the ID parameter using proper HTML encoding
• Restrict network access to the vulnerable endpoints to trusted users only
• Consider disabling the share functionality entirely if it is not critical to operations

# Example WAF rule for blocking HTML injection attempts (ModSecurity syntax)
SecRule ARGS:id "@rx <[^>]*>" \
    "id:1001,\
    phase:2,\
    block,\
    msg:'Potential HTML Injection in ID parameter',\
    log,\
    severity:'MEDIUM'"