CVE-2024-2288: Lollms Web UI CSRF Vulnerability

CVE-2024-2288 Overview

CVE-2024-2288 is a Cross-Site Request Forgery (CSRF) vulnerability in the parisneo/lollms-webui repository, affecting Lollms versions up to and including 7.3.0. The flaw resides in the profile picture upload functionality and allows attackers to change a victim's profile picture without consent. Attackers can leverage the same endpoint to flood the filesystem with arbitrary files, producing a denial-of-service condition. The vulnerability also enables stored cross-site scripting (XSS), letting attackers execute arbitrary JavaScript in the victim's browser session. The maintainers resolved the issue in Lollms version 9.3.

Critical Impact
Attackers can trigger unauthorized profile picture uploads through forged requests, leading to stored XSS execution and filesystem-level denial of service against Lollms Web UI instances.

Affected Products

Lollms Web UI versions up to and including 7.3.0
parisneo/lollms-webui repository builds prior to 9.3
Deployments using the profile picture upload endpoint

Discovery Timeline

2024-06-06 - CVE-2024-2288 published to NVD
2026-06-17 - Last updated in NVD database

Technical Details for CVE-2024-2288

Vulnerability Analysis

The Lollms Web UI exposes a profile picture upload endpoint that lacks anti-CSRF protection [CWE-352]. The endpoint accepts state-changing POST requests without validating an origin token, referer header, or SameSite cookie restriction. An attacker who entices an authenticated Lollms user to visit a malicious page can submit a cross-origin upload request that the browser will execute with the victim's session cookies.

The upload handler accepts attacker-controlled file content and writes it to disk under the user's profile context. Because the application does not sanitize the rendered output of profile assets, attackers can craft a payload that executes JavaScript when other users view the affected profile. This converts a forced upload into a stored XSS primitive.

The same endpoint also lacks rate limiting and file count controls. Repeated forged uploads write attacker-controlled files to the server, exhausting disk capacity and degrading service availability.

Root Cause

The root cause is the absence of CSRF token validation on the profile picture upload route in parisneo/lollms-webui releases up to 7.3.0. The handler trusts that any authenticated request reaching it is intentional, omitting verification of request origin or a synchronized token. Insufficient validation of uploaded file content compounds the issue by permitting executable script payloads to persist.

Attack Vector

Exploitation requires user interaction. The attacker hosts a page containing a hidden form or fetch call targeting the Lollms upload endpoint. When an authenticated Lollms user loads the page, the browser submits the request with valid session cookies. The server processes the upload as legitimate, replacing the victim's profile image with attacker-supplied content. Repeated automated submissions amplify the impact into a denial-of-service against the filesystem, while script-bearing payloads achieve stored XSS when rendered to other users.

No public proof-of-concept is published, and the vulnerability is not listed in the CISA Known Exploited Vulnerabilities catalog. See the Huntr Bounty Information for additional technical context.

Detection Methods for CVE-2024-2288

Indicators of Compromise

Unexpected modifications to user profile picture assets within the Lollms data directory
High volumes of POST requests to the profile picture upload endpoint originating from a single session within a short window
Profile image files containing HTML or JavaScript markup rather than valid image binary headers
Inbound requests to the upload endpoint with Origin or Referer headers pointing to untrusted domains

Detection Strategies

Inspect web server access logs for upload requests where the Referer header does not match the Lollms application domain
Monitor the Lollms uploads directory for sudden growth in file count or total size
Validate stored profile assets against expected image MIME signatures and flag files that contain script tags

Monitoring Recommendations

Alert on anomalous upload frequency per authenticated user against a baseline
Track filesystem disk utilization on hosts running Lollms Web UI and alert on rapid consumption
Log and review all responses from the profile picture endpoint for HTTP status anomalies and unexpected content types

How to Mitigate CVE-2024-2288

Immediate Actions Required

Upgrade parisneo/lollms-webui to version 9.3 or later, which contains the fix in commit ed085e6effab2b1e25ba2b00366a16ff67d8551b
Restrict access to the Lollms Web UI to trusted networks until the upgrade is complete
Audit existing profile picture assets for embedded script content and remove any suspicious files
Force re-authentication and invalidate active sessions after patching

Patch Information

The vendor resolved CVE-2024-2288 in Lollms Web UI version 9.3. Review the fix in the GitHub Commit Update. The patch introduces request validation on the profile picture upload route and addresses the stored XSS vector.

Workarounds

Place the Lollms Web UI behind a reverse proxy that enforces Origin and Referer header checks on POST requests to upload endpoints
Configure session cookies with the SameSite=Strict attribute to block cross-origin credentialed requests
Apply Content Security Policy (CSP) headers restricting script execution sources to mitigate XSS impact
Enforce filesystem quotas on the Lollms uploads directory to limit denial-of-service potential

bash

# Example nginx configuration enforcing SameSite cookies and Referer validation
proxy_cookie_path / "/; SameSite=Strict; Secure; HttpOnly";

location /upload {
    if ($http_referer !~* ^https://lollms\.example\.com/) {
        return 403;
    }
    proxy_pass http://lollms_backend;
}