CVE-2026-1115: parisneo/lollms Stored XSS

Overview

A critical stored Cross-Site Scripting (XSS) vulnerability, CVE-2026-1115, was identified in the parisneo/lollms application. The flaw resides in the platform’s social feature, specifically within the create_post function. This function failed to sanitize user input before storing it in the database, allowing malicious JavaScript code to be permanently embedded.

Technical Details

The vulnerability is located in backend/routers/social/__init__.py. When a user creates a post, the content is directly assigned to the DBPost model without any validation or sanitization. This stored payload is then rendered and executed in the browsers of any user who views the compromised post in the Home Feed. The attack requires no special privileges, and the victim only needs to view the malicious content, making exploitation straightforward.

Impact and Risks

With a CVSS score of 9.6, this vulnerability poses a severe risk. An attacker can craft a post containing malicious JavaScript. When other users, including administrators, view this post, the script executes in their browser session. This can lead directly to:

• Account takeover through session hijacking.
• Unauthorized actions performed on behalf of the victim.
• Wormable attacks where the malicious post spreads automatically. Given the social nature of the feature, a single compromised post could rapidly affect numerous users on the platform. For more on how such exploits can lead to data exposure, see our breach reports.

Remediation

The issue is fully resolved in version 2.2.0 of parisneo/lollms. The only complete mitigation is to upgrade all affected instances to this version immediately. There is no effective workaround; disabling the social feature would be required to block the attack vector, which may not be practical for all deployments.

Security Insight

This vulnerability highlights the persistent risk of XSS in modern web applications, especially in user-generated content features that are often rushed to market. It mirrors a common pattern seen in other platforms where input sanitization is an afterthought rather than a core design principle. The high severity underscores that even non-RCE flaws in interactive components can be critically damaging, enabling lateral movement and privilege escalation within an application. Stay updated on similar threats by following our security news.

Further Reading

• All Critical Vulnerabilities
• Recent Critical Vulnerabilities
• Top Critical CVEs