CVE-2026-121255 in Chrome
Summary
by MITRE • 11/03/2026
Incorrect authorization in Autofill AI in Google Chrome prior to 157.0.8103.38 allowed a remote attacker to obtain cross-origin data via a crafted HTML page. (Chromium security severity: Medium)
If you want to get best quality of vulnerability data, you may have to visit VulDB.
Analysis
by VulDB Data Team • 11/03/2026
The vulnerability identified as an incorrect authorization flaw within the Autofill AI component of Google Chrome prior to version 157.0.8103.38 represents a significant breach in cross-origin isolation mechanisms. This issue stems from improper access control checks that fail to adequately restrict data retrieval operations based on the origin context of the requesting entity. In modern web browsers, strict separation between origins is fundamental to maintaining user privacy and preventing malicious sites from accessing sensitive information belonging to other domains. The Autofill AI subsystem, which assists users by predicting and filling in form fields with stored personal data such as addresses or credit card numbers, must operate within a tightly controlled security boundary. When this boundary is compromised due to insufficient authorization verification, the browser inadvertently exposes internal state or user-specific data to contexts that should not have access to it.
From a technical perspective, the flaw allows a remote attacker to exploit the misconfigured permissions by crafting an HTML page designed to trigger specific API calls within the Autofill AI module. By leveraging standard web technologies such as JavaScript and DOM manipulation, the malicious page can initiate requests that bypass origin checks. This enables the extraction of cross-origin data, which may include personally identifiable information stored locally in the browser profile. The severity is classified as Medium by Chromium security standards because while it does not typically allow for arbitrary code execution or full system compromise, it facilitates a direct violation of user privacy and confidentiality. An attacker could potentially harvest sensitive form inputs that were previously protected from cross-site access, leading to unauthorized data disclosure without requiring physical access to the victim's machine or complex social engineering beyond hosting the malicious page.
The operational impact of this vulnerability is primarily centered on information leakage and potential identity theft scenarios. If a user visits a compromised website while logged into services where autofill data is enabled for sensitive fields, the attacker can programmatically retrieve that stored information. This could include email addresses, phone numbers, physical addresses, or even partial payment details depending on how the browser stores and exposes these values to web APIs. Such exposure undermines trust in the browser's security model and violates principles of least privilege by granting a low-privilege context access to high-sensitivity local data. Furthermore, this type of vulnerability can serve as an initial vector for more sophisticated attacks, such as session hijacking if combined with other exploits or used to gather intelligence for targeted phishing campaigns against specific individuals based on their stored personal details.
To mitigate the risks associated with this authorization flaw, users must upgrade Google Chrome to version 157.0.8103.38 or later immediately upon availability. This update contains patches that enforce stricter origin checks within the Autofill AI subsystem, ensuring that data retrieval operations are properly validated against the requesting context's permissions. In environments where automatic updates cannot be deployed instantly, administrators should consider restricting access to untrusted websites through enterprise policies and disabling autofill features for sensitive fields until the patch is applied. Additionally, users can enhance their security posture by regularly reviewing browser extensions and ensuring that no malicious scripts are injected into browsing sessions via compromised plugins.
This vulnerability aligns with CWE-269, which describes Improper Authorization, as it involves a failure to enforce proper access controls on system resources. It also relates to CWE-359, the Exposure of Private Information to an Unauthorized Actor, highlighting the direct consequence of the flaw. From an offensive security perspective, this behavior is consistent with techniques found in the MITRE ATT&CK framework under T1005, which covers Data from Local System Retrieval. Understanding these mappings helps security teams categorize similar incidents and apply appropriate detection rules to identify potential exploitation attempts involving unauthorized data access through browser-based mechanisms.