macOS CVE-2026-28827

Local attacker with ability to execute an app on macOS Sequoia (before 15.7.5), Sonoma (before 14.8.5), or Tahoe (before 26.4). The vulnerable app exploits a path traversal flaw to escape sandbox restrictions. No user interaction required.

This vulnerability carries significant real-world risk despite the absence of published CVSS, CVSS vector, and EPSS scores. Sandbox escape vulnerabilities are historically prioritized by attackers and security researchers because they represent a fundamental breach of macOS security architecture. The fact that Apple released patches across three major macOS versions (Sequoia 15.7.5, Sonoma 14.8.5, and Tahoe 26.4) indicates widespread impact and demonstrates vendor recognition of severity. Without KEV status confirmation or public proof-of-concept details from the references provided, the actual exploitation likelihood in the wild remains uncertain; however, the technical impact of sandbox escape is universally high—an attacker who escapes the sandbox gains access to protected system resources, user files, and inter-process communication mechanisms. This vulnerability should be treated as a priority patch target using the SSVC framework due to its technical impact (sandbox bypass = high), likelihood of automatable exploitation once public details emerge, and potential for active exploitation in targeted attacks. Organizations should prioritize patching over other lower-impact vulnerabilities.

An attacker distributes a trojanized application through unofficial channels that, once installed and granted basic permissions by the user, exploits the path validation flaw to escape the sandbox. The malicious app could then enumerate the user's file system, exfiltrate sensitive documents, access keychain credentials, or install persistence mechanisms without triggering sandbox alerts. The attack requires user interaction to install the app but no special privileges, as the vulnerability exists in the kernel-level sandbox enforcement mechanism rather than in individual application logic. Given the parsing nature of the flaw, exploitation likely involves crafting specially-formed file paths at runtime to deceive the sandbox validation layer, making the attack difficult to detect through static analysis alone.