CVE-2026-0533: Autodesk Fusion Stored XSS Vulnerability

CVE-2026-0533 Overview

A Stored Cross-Site Scripting (XSS) vulnerability has been identified in the Autodesk Fusion desktop application. The vulnerability occurs when a maliciously crafted HTML payload is embedded in a design name, which is then rendered during the delete confirmation dialog. When a user clicks on the affected element, the injected script executes within the context of the application, potentially allowing an attacker to read local files or execute arbitrary code.

Critical Impact

This vulnerability enables attackers to execute arbitrary code in the context of the current process, potentially leading to local file access and system compromise through a trusted desktop application.

Affected Products

• Autodesk Fusion Desktop Application (Windows)
• Autodesk Fusion Desktop Application (macOS)

Discovery Timeline

• 2026-01-22 - CVE CVE-2026-0533 published to NVD
• 2026-01-22 - Last updated in NVD database

Technical Details for CVE-2026-0533

Vulnerability Analysis

This vulnerability represents a classic Stored XSS attack vector (CWE-79) that leverages improper neutralization of user input during web page generation. The Autodesk Fusion desktop application fails to properly sanitize HTML content within design names before rendering them in the user interface. When the delete confirmation dialog displays a maliciously crafted design name, the embedded HTML payload is interpreted as executable code rather than being treated as plaintext.

The local attack vector requires user interaction—specifically, the victim must click on the compromised element within the delete confirmation dialog. Once triggered, the attacker gains the ability to execute scripts within the application's context, which can lead to unauthorized file system access and arbitrary code execution with the privileges of the running process.

Root Cause

The root cause of this vulnerability lies in insufficient input validation and output encoding within the Autodesk Fusion application. Design names are accepted without proper sanitization, allowing HTML and JavaScript content to be stored. Subsequently, when these names are rendered in UI elements such as the delete confirmation dialog, the application fails to encode the output, causing the browser engine or embedded web view to interpret the malicious payload as executable code.

Attack Vector

The attack requires local access and follows a multi-stage approach:

1. Payload Injection: An attacker creates or modifies a design with a malicious HTML/JavaScript payload embedded in the design name
2. Social Engineering: The victim user is induced to delete the affected design, triggering the delete confirmation dialog
3. Payload Execution: When the dialog renders and the user clicks on the affected element, the malicious script executes
4. Impact Realization: The attacker's code can access local files, exfiltrate data, or execute additional arbitrary commands within the application context

The vulnerability exploits the trust relationship between the desktop application and its embedded web rendering component, allowing scripts to escape the typical browser sandbox restrictions.

Detection Methods for CVE-2026-0533

Indicators of Compromise

• Unusual design names containing HTML tags such as <script>, <img>, <iframe>, or event handlers like onclick, onerror
• Unexpected network connections originating from the Autodesk Fusion process
• Evidence of local file access attempts from JavaScript contexts within the application
• Anomalous process spawning from Autodesk Fusion application

Detection Strategies

• Monitor for design names containing suspicious HTML or JavaScript patterns using content inspection rules
• Implement endpoint detection rules to identify script execution anomalies within the Fusion application context
• Configure SentinelOne Singularity to detect process injection and unusual child process creation from Fusion360.exe or Autodesk Fusion processes
• Review application logs for unusual rendering errors or script execution warnings

Monitoring Recommendations

• Enable verbose logging for Autodesk Fusion application events
• Configure SentinelOne behavioral AI to monitor for file system access patterns indicative of data exfiltration
• Establish baseline behavior for Fusion application network activity and alert on deviations
• Monitor for attempts to access sensitive local files from web rendering contexts

How to Mitigate CVE-2026-0533

Immediate Actions Required

• Update Autodesk Fusion to the latest patched version immediately
• Review existing designs for suspicious or malformed names containing HTML content
• Restrict design import from untrusted sources until the patch is applied
• Enable SentinelOne's application control features to monitor Fusion application behavior

Patch Information

Autodesk has released a security update addressing this vulnerability. Users should download the latest version from the official Autodesk distribution channels:

• Windows: Download the updated Fusion Client Downloader for Windows
• macOS: Download the updated Fusion Client Downloader for macOS

For complete details, refer to the Autodesk Security Advisory ADSK-SA-2026-0001.

Workarounds

• Exercise caution when opening designs from untrusted sources or collaborators
• Avoid clicking on elements within delete confirmation dialogs for designs with unusual or suspicious names
• Implement network segmentation to limit potential data exfiltration paths
• Use SentinelOne's network control capabilities to restrict unauthorized outbound connections from the Fusion application

# SentinelOne Deep Visibility query to detect suspicious Fusion activity
# Monitor for unusual script execution patterns
EventType = "Process" AND 
ProcessName CONTAINS "Fusion" AND 
(CommandLine CONTAINS "script" OR CommandLine CONTAINS "javascript")