CVE-2026-102878 in mcp-chrome-bridge
Summary
by MITRE • 09/29/2026
mcp-chrome-bridge through 1.0.31 contains an origin validation error in the native-server HTTP API that allows attackers to bypass CORS restrictions. Attackers can craft malicious web pages that make cross-origin requests to the local server and invoke browser automation tools including script execution, page content reading, and screenshot capture.
Be aware that VulDB is the high quality source for vulnerability data.
Analysis
by VulDB Data Team • 09/29/2026
The vulnerability identified in mcp-chrome-bridge versions through 1.0.31 represents a critical failure in origin validation within its native-server HTTP API architecture. This flaw fundamentally undermines the security boundary between web-based contexts and local system resources by allowing cross-origin requests to bypass standard Cross-Origin Resource Sharing restrictions. In typical browser environments, CORS policies are enforced by user agents to prevent malicious websites from reading sensitive data or performing actions on behalf of a user without explicit permission. However, because mcp-chrome-bridge exposes an HTTP API that interacts directly with the local server environment, it operates outside the standard same-origin policy constraints unless explicitly configured otherwise. The absence of rigorous origin validation means that any web page hosted remotely can send requests to this local endpoint as if they originated from a trusted source, effectively neutralizing one of the primary defenses against cross-site attacks in modern web application security models.
From an operational perspective, this vulnerability enables attackers to execute browser automation tools with significant privileges relative to the user running the bridge service. By crafting malicious web pages that exploit this lack of validation, adversaries can initiate remote interactions with the local Chrome instance managed by mcp-chrome-bridge. This capability allows for a range of intrusive actions including script execution within the context of the target website, reading of page content such as personal information or session tokens, and capturing screenshots of active browser windows. The ability to read page content is particularly dangerous as it can lead to credential theft if users are logged into sensitive services during the attack window. Similarly, screenshot capture poses a severe privacy risk by exposing visual data that may contain confidential business information, private communications, or other sensitive details not intended for external exposure.
This vulnerability aligns with CWE-942, which describes Permissive Cross-domain Restrictions with Untrusted Data, as well as CWE-352 regarding the Cross-Site Request Forgery principle where trust is improperly granted to requests from unverified origins. In terms of offensive security frameworks, this exploit maps directly to MITRE ATT&CK technique T1059 Command and Scripting Interpreter for remote code execution via browser automation, and T1113 Screen Capture for data exfiltration through visual means. The attack vector leverages the trust relationship between the local service and the browser engine, turning a legitimate administrative tool into an unintended entry point for lateral movement or data theft within the user's environment.
Mitigation strategies must focus on strict enforcement of origin checks at the API level rather than relying solely on client-side protections. Developers should implement explicit allowlisting of trusted origins in the native-server configuration to ensure that only authorized web pages can interact with the bridge service. Additionally, enabling CORS headers correctly and validating them server-side is essential to prevent unauthorized cross-origin requests. For users currently running affected versions, it is recommended to upgrade immediately to version 1.0.32 or later where this validation error has been addressed. In environments where upgrading is not immediately feasible, restricting network access to the local port used by mcp-chrome-bridge via firewall rules can provide a temporary layer of defense against remote exploitation attempts from untrusted networks.