CVE-2026-12228
MediumCVSS 5.4Exploitation Probability (EPSS)
Low risk13th percentile - higher than 13% of all known CVEs
Summary
Stored XSS vulnerability in the POST /api/prompts/share endpoint in parisneo/lollms (latest version). An attacker can store malicious JavaScript in a message, which executes in the victim's browser when opened, enabling data theft and account takeover.
Risk Assessment
Risk of arbitrary JavaScript execution in the victim's context, which could lead to session theft, data theft, and account takeover.
Recommendation
Apply proper server-side input sanitization. Update lollms to a patched version if available. Restrict message sending to trusted users.
Related vulnerabilities
- CVE-2026-110812Critical
Deserialization of untrusted data in the Gremlin graph data import component in the Amazon Neptune engine. An authenticated remote actor with standard database permissions can execute arbitrary code and access internal service credentials.
- CVE-2026-57155Critical
In OPNsense prior to 26.1.11, the Firewall Alias GeoIP importer uses an unvalidated CSV country_code value to construct output filenames, allowing path traversal. A user with Firewall Alias Edit privilege can supply a crafted archive and write attacker-controlled content as root, then leverage cron to execute commands as root.
- CVE-2026-110958Critical
A heap-based buffer overflow exists in Wavlink WN530HG4B (M30HG4B_V260724) in the file api.cgi. The attack can be launched remotely and the exploit has been made public.
- CVE-2026-105474Critical
A server-side template injection (SSTI) vulnerability exists in SolusVM 2 up to version 2.0.0-44576, allowing remote authenticated users to execute arbitrary code as root.
- CVE-2026-47168Critical
In EspoCRM prior to 9.3.7, an authenticated administrator can supply crafted serialized PHP object data stored in MassAction or Export parameters, which is later deserialized without class restrictions, leading to arbitrary code execution.
- CVE-2026-110957Critical
In Wavlink WL-WN530BE2-A up to M30BE2_V250604, the function sub_12580 in api.cgi (Login Processing component) is vulnerable to OS command injection via the REMOTE_ADDR argument. The attack can be launched remotely and the exploit has been disclosed publicly.
- CVE-2026-99999Critical
In the Linux kernel, a vulnerability in the IPv4 stack was resolved: if a packet with bad checksum metadata enters the stack, skb_checksum_help can corrupt the network header. Additional validation of checksum_start before completing the checksum was added.
- CVE-2026-109892Critical
Incorrect Privilege Assignment vulnerability in Shabti Kaplan Frontend Admin by DynamiApps (acf-frontend-form-element) allows Privilege Escalation. This issue affects Frontend Admin by DynamiApps from n/a through 3.29.13.
- CVE-2026-109698Critical
Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability in Kitae Park Mang Board WP (mangboard) allows SQL Injection. This issue affects Mang Board WP from n/a through 2.4.4.
- CVE-2026-109688Critical
Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability in Link Whisper INC Link Whisper Free (link-whisper) allows Blind SQL Injection. This issue affects Link Whisper Free from n/a through 0.9.7.
Original NVD description (English source)
A stored cross-site scripting (XSS) vulnerability exists in the `POST /api/prompts/share` endpoint of parisneo/lollms (latest version). The endpoint stores attacker-controlled `prompt_content` into `DBDirectMessage.content` without server-side sanitization. When a victim opens the direct message (DM) thread, the message is rendered by the DM UI through `MessageContentRenderer`, which uses `v-html` to insert rendered HTML into the DOM. The frontend sanitizer, which is regex-based, fails to comprehensively sanitize attacker-controlled HTML, allowing malicious payloads to execute in the victim's browser context. This vulnerability enables any authenticated user to send a malicious prompt-share message to another user's inbox, leading to arbitrary JavaScript execution, authenticated actions as the victim, exposure of same-origin application data, and potential account takeover.
Vulnerability data from NVD (NIST) · CISA KEV · EPSS