Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB
CVE-2026-106102 Overview
CVE-2026-106102 is a cross-site scripting (XSS) vulnerability in Quasar Framework, a Vue.js framework for building high-performance user interfaces. Versions prior to 2.22.0 contain a server-side rendering (SSR) flaw in the getHead() serializer located in ui/src/plugins/meta/Meta.js. The serializer uses getAttr() to interpolate values from useMeta() into title, meta, link, and script markup without HTML text or quoted-attribute encoding. The injectServerMeta() function then appends that output to the raw server-rendered response. Attackers who control dynamic page metadata can break out of the intended HTML context and inject executable markup before hydration.
Critical Impact
Unauthenticated attackers can inject arbitrary HTML and JavaScript into SSR responses through fields such as post titles, product names, or display names, leading to full client-side compromise.
Affected Products
- Quasar Framework versions prior to 2.22.0
- Applications using Quasar SSR mode with useMeta() composable
- Any Quasar SSR deployment interpolating user-controlled values into page metadata
Discovery Timeline
- 2026-10-06 - CVE CVE-2026-106102 published to NVD
- 2026-10-06 - Last updated in NVD database
Technical Details for CVE-2026-106102
Vulnerability Analysis
The flaw [CWE-79] originates in Quasar's SSR meta plugin. During server rendering, getHead() builds <title>, <meta>, <link>, and <script> tags by concatenating values supplied through the useMeta() composable. Because getAttr() performed no HTML entity encoding and no attribute-value quoting enforcement, any character that terminates an attribute or tag context passed through unmodified. The output was then written directly into the HTML stream by injectServerMeta().
An attacker who influences metadata fields—such as a blog post title stored in a database or a product name rendered on a catalog page—can insert characters like ", >, or </script> to escape the intended context. Injected markup executes in the user's browser before Vue hydration completes, giving the attacker a reliable XSS primitive against every visitor to the affected page.
The client-side apply() path is unaffected because it uses DOM APIs that encode attribute values automatically. Only the SSR serialization path is vulnerable.
Root Cause
The root cause is missing output encoding in the SSR meta serializer. getAttr() concatenated attribute values directly into the HTML template without applying HTML entity encoding to characters such as &, <, >, ", and ', and without encoding dangerous Unicode sequences inside JSON contexts.
Attack Vector
Exploitation requires no authentication or user interaction beyond visiting the affected page. An attacker submits crafted content through any application input that eventually flows into useMeta() fields. When the server renders the page, the malicious payload is embedded in the response HTML and executed by the victim's browser.
// Patch applied in ui/src/plugins/meta/Meta.js (version 2.22.0)
const encodeHtmlChars = {
'&': '&',
'<': '<',
'>': '>',
'"': '"',
"'": '''
}
const encodeJsonChars = {
'<': String.raw`\\u003C`,
'>': String.raw`\\u003E`,
'&': String.raw`\\u0026`,
'\\u2028': String.raw`\\u2028`,
'\\u2029': String.raw`\\u2029`
}
const validAttrName = /^[^\s"'>/=]+$/
function isValidAttrName(name) {
return (
validAttrName.test(name) && name.includes(String.fromCodePoint(0)) === false
)
}
function encodeHtml(value) {
return String(value).replaceAll(/[&<>"']/g, char => encodeHtmlChars[char])
}
Source: GitHub Commit 11505af
Detection Methods for CVE-2026-106102
Indicators of Compromise
- Unexpected <script> tags or inline event handlers appearing inside <head> elements in server-rendered HTML responses.
- Metadata fields in the application database containing HTML control characters such as <, >, ", or </script> fragments.
- Outbound requests from client browsers to unfamiliar domains shortly after visiting pages that render user-supplied titles or descriptions.
Detection Strategies
- Inventory applications using Quasar Framework and verify the installed version against 2.22.0 using npm ls quasar or lockfile inspection.
- Review all call sites of useMeta() and identify which fields receive user-controlled or externally sourced values.
- Scan historical SSR response logs for metadata values containing unescaped HTML or quote characters.
Monitoring Recommendations
- Deploy content security policy (CSP) reporting to surface inline script execution originating from <head> content.
- Monitor web application firewall (WAF) logs for payloads targeting title, description, and Open Graph fields.
- Alert on database writes to metadata columns that contain HTML tag characters or JavaScript protocol handlers.
How to Mitigate CVE-2026-106102
Immediate Actions Required
- Upgrade Quasar Framework to version 2.22.0 or later across all SSR deployments.
- Audit all user-controlled inputs that flow into useMeta() and apply server-side sanitization before storage.
- Enforce a strict Content-Security-Policy header that blocks inline scripts and restricts script sources to known origins.
Patch Information
The fix is included in Quasar Framework Release v2.22.0. The patch introduces encodeHtml() and encodeJsonChars helpers and validates attribute names against a strict character set. Full technical details are documented in GitHub Security Advisory GHSA-pq96-jpmf-w254.
Workarounds
- Pre-encode HTML-significant characters in metadata values before passing them to useMeta() if an immediate upgrade is not feasible.
- Disable SSR for routes that render user-controlled metadata and fall back to client-side rendering, which uses safe DOM APIs.
- Apply a WAF rule that rejects requests whose metadata-bound parameters contain <, >, or quote characters.
# Upgrade Quasar Framework to the patched release
npm install quasar@^2.22.0
# Verify the installed version
npm ls quasar
Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.
Sources: National Vulnerability Database, CISA KEV, Exploit-DB

