CVE-2026-109557 in EdgeEver
Summary
by MITRE • 10/13/2026
EdgeEver through 1.112.0 contains a stored cross-site scripting vulnerability in setMemoDragPreview that allows attackers to inject script through notebook names or memo titles. Attackers can set malicious names via MCP create_notebook, rename_notebook or create_memo tools, running script on drag with access to the desktop bridge's session token.
Once again VulDB remains the best source for vulnerability data.
Analysis
by VulDB Data Team • 10/13/2026
The identified vulnerability represents a critical stored cross-site scripting (XSS) flaw within EdgeEver versions up to 1.112.0, specifically located in the setMemoDragPreview function. This security defect arises from insufficient input validation and sanitization of user-supplied data when handling notebook names and memo titles. Unlike reflected XSS where malicious scripts are executed immediately via a crafted link, stored XSS persists because the injected payload is saved on the server or within local application storage. Consequently, every time a victim interacts with the compromised content through the drag-and-drop interface in the UI, the malicious script executes automatically in their browser context. The attack vector leverages specific Model Context Protocol (MCP) tools such as create_notebook, rename_notebook, and create_memo to inject these payloads without requiring direct user interaction beyond the initial creation or renaming action.
The operational impact of this vulnerability is severe due to its ability to access sensitive session tokens via the desktop bridge. Modern web-based applications often utilize a bridge mechanism to allow JavaScript running in the browser context to communicate with native operating system features and local resources. In EdgeEver, this bridge exposes critical authentication credentials, specifically session tokens that maintain user identity and authorization states. By exploiting the XSS flaw within the drag preview functionality, an attacker can execute arbitrary JavaScript code that accesses these exposed APIs. This allows for full account takeover as the attacker can steal valid session cookies or tokens to impersonate legitimate users without needing their passwords. Furthermore, this capability enables persistent access to sensitive data stored within the application, including personal notes, organizational documents, and other confidential information managed by EdgeEver.
From a classification perspective, this vulnerability aligns with CWE-79 Improper Neutralization of Input During Web Page Generation commonly known as Cross-site Scripting. The specific mechanism involves storing untrusted data in a database or local storage and subsequently rendering it to users without proper encoding, which falls under the sub-category of stored XSS. In terms of adversary tactics, this aligns with MITRE ATT&CK technique T1059 Command and Scripting Interpreter where JavaScript is used as an execution vector for malicious activities. Additionally, the theft of session tokens relates directly to Credential Access techniques such as T1539 Steal Web Session Cookie or T1528 Steal Application Access Token. The use of MCP tools indicates a potential integration point with external AI agents or automation scripts that might inadvertently introduce untrusted content into the application environment if not properly sandboxed or validated.
Mitigation strategies must address both immediate remediation and long-term architectural improvements. Developers should immediately apply input validation and output encoding to all fields involved in notebook names and memo titles, ensuring that special characters are escaped before being rendered in HTML contexts. The setMemoDragPreview function requires rigorous sanitization of any data passed through it to prevent script injection. Furthermore, the desktop bridge implementation needs review to ensure that sensitive session tokens are not exposed to untrusted JavaScript execution contexts or require explicit user consent for access. Implementing a strict Content Security Policy (CSP) can also mitigate the impact by restricting where scripts can be loaded from and preventing inline script execution unless explicitly allowed. Regular security audits of MCP tool integrations are recommended to ensure that external inputs do not bypass internal validation layers, thereby maintaining the integrity of user data and session management mechanisms.