[Stored XSS, CVE] Seafile SVG

While exploring secure image handling in web applications, I came across a stored cross-site scripting (XSS) vulnerability in Seafile, a popular open-source file storage and collaboration platform.



This vulnerability allows an attacker to upload a malicious SVG file that executes JavaScript when viewed directly or embedded, potentially compromising user sessions and data.

Summary

Seafile does not adequately sanitize SVG files uploaded by users, allowing arbitrary JavaScript execution when the SVG is accessed via its raw URL. This issue is made more dangerous because the raw vulnerable URL is exposed in the page source and can be shared with other users, leading to widespread XSS.

Proof of Concept (PoC)

1. Log in to a Seafile instance.

2. Upload an SVG file containing embedded JavaScript (e.g., <svg><script>alert(document.domain)</script></svg>).

3. Right-click the uploaded image and select “Open in new tab.”

4. The raw SVG URL opens and the JavaScript payload executes.

5. Additionally, inspecting the page source of the file preview reveals the same raw vulnerable URL. If this file is shared, anyone with access can trigger the XSS.


Impact

· Stored XSS affecting all users who view or open the malicious file.

· Possible session hijacking, account takeover, or actions performed on behalf of the user.

· Risk increases in shared environments (teams, organizations).


Bounty & Recognition 



This vulnerability was acknowledged and rewarded with a $200 bounty through the responsible disclosure program.

CVE: CVE-2025-65516

Comments

Post a Comment

Popular posts from this blog

[XSS] Breaking ‘safe’ embeds via frame-src bypass

Image
Overview We pat ourselves on the back when our WAF blocks a classic <script> alert. But modern web applications are a tapestry of first-party code and third-party services. What happens when an attacker doesn't need to break your walls, but simply uses a door you left open for a trusted guest? The Vulnerability Let's set the scene. Your application, trusted-site.com, has secure server-side filters. It correctly neutralizes or blocks any user input containing <script>, onerror=, or other obvious XSS vectors. However, to support rich content, you allow certain HTML tags like <iframe> and <object>. These are necessary for embedding videos, maps, or other third-party widgets. An attacker discovers this policy. They don't try to inject JavaScript directly. Instead, they inject a gateway: <iframe src="https://vulnerable-widget-provider.com/page-with-xss.html" style="position:fixed;top:0;left:0;width:100vw;height:100vh;border:none;"...
Read more

About Me

My interest in cybersecurity started when I was a kid, though I didn't fully realize it at the time. I spent a lot of time playing online games and would often run into cheaters. It was frustrating, but it also made me curious. How were they doing it? My research led me to discover that cheats were often made with code, particularly in Java. This sparked my interest, and soon I was experimenting with Cheat Engine. Seeing how things worked behind the scenes was fascinating and made me want to learn more about how computers and software functioned. My exploration eventually led me to web development, but cybersecurity always grabbed my attention the most. When I was around 8 or 9, I noticed a security weakness on my school's website involving easily guessable credentials. Gaining that firsthand view of how predictable information could be exploited felt like my first lesson in security and social engineering. That moment pushed me to learn even more. I started digging into diffe...
Read more

[BYPASS] IKEv2 Fortinet Misconfiguration

Image
Overview The target was a Fortinet Next-Generation Firewall (NGFW), a formidable system relying on aggressive Deep Packet Inspection (DPI). Its directive was simple: block all unauthorized tunneling. My objective was proving that its "threat intelligence" had exploitable policy blind spots. Initial attempts with solutions like OpenVPN and WireGuard failed instantly. This was not simple port blocking. The FortiGate performed protocol fingerprinting, immediately recognizing the distinct traffic signatures of these protocols and dropping the connections. Any consumer VPN signature was a red flag. The challenge was confirming and exploiting a predictable policy exclusion. The goal crystallized: Protocol-Based Bypass. The IKEv2 Differential Testing revealed the key vulnerability: IKEv2 (Internet Key Exchange version 2) traffic showed inconsistent signs of life while other protocols were instantly terminated. IKEv2 is part of the IPsec suite and is specifically favored for enterpri...
Read more