Executive Summary

This vulnerability affects DeepChat, an open-source AI agent platform. Before version 1.0.4-beta.1, there was an incomplete fix for a previous vulnerability (CVE-2025-55733) that left DeepChat open to arbitrary protocol execution bypass, which can lead to remote code execution (RCE).

The issue arises because while the patch restricted the api.openExternal() function in the renderer's preload script, it failed to sanitize native Electron pop-up window handlers. An attacker or a compromised AI endpoint can return a Markdown link that triggers a native window interception with target="_blank" in the tabPresenter.ts file. This interception forwards the malicious URL directly to shell.openExternal(url), bypassing the security check isValidExternalUrl, allowing execution of arbitrary protocols.

This vulnerability is fixed in DeepChat version 1.0.4-beta.1.

Useful 0 Not Useful 0

Impact Analysis

This vulnerability can have severe impacts because it allows an attacker or compromised AI endpoint to execute arbitrary protocols on the target system remotely.

• Remote Code Execution (RCE) - attackers can run malicious code on the victim's machine.
• Complete bypass of security boundaries designed to validate external URLs.
• Potential compromise of confidentiality, integrity, and availability of the affected system.

The CVSS v3.1 base score of 9.6 indicates a critical severity with network attack vector, low attack complexity, no privileges required, user interaction required, and high impact on confidentiality, integrity, and availability.

Useful 0 Not Useful 0

Mitigation Strategies

To mitigate this vulnerability, upgrade DeepChat to version 1.0.4-beta.1 or later, where the issue with arbitrary protocol execution bypass has been fixed.

The patch restricts api.openExternal() inside the renderer's preload/index.ts script and properly sanitizes native Electron pop-up window handlers to prevent malicious URLs from bypassing security boundaries.

Useful 0 Not Useful 0