CVE-2026-28900

CVE-2026-28900 is a Gatekeeper bypass via quarantine metadata stripping during ZIP archive extraction. The CVSS 5.5 score fundamentally misrepresents the risk. Gatekeeper is the final discretionary checkpoint before arbitrary code executes on macOS — bypassing it doesn't simply 'allow a suspicious file to run,' it removes the gate entirely, converting a socially acceptable vector (a ZIP file) into a code execution primitive. The CVSS scoring model treats this as a file-integrity issue rather than an execution issue, which fails to capture attack-chain composition. The patch description mentions 'additional checks,' and this phrasing is analytically significant. It indicates Apple patched the symptom — this specific ZIP interaction — rather than redesigning how archive utilities propagate quarantine metadata. This is the seventh documented Gatekeeper bypass via archive attribute stripping in seven years (CVE-2017-2389, CVE-2018-4168, CVE-2019-8513, CVE-2020-10015, CVE-2021-1800, and now this). Each fix has been format-specific, each has worked until a new archive format or manipulation technique emerged, and each arrived in a point release without architectural redesign. This is a known recurrence pattern, not an anomaly. What this means in practice: if you manage macOS endpoints, treat this as a high-severity risk regardless of the CVSS score. The vulnerability creates a reliable primitive that chains with Safari exploits, social engineering, or other delivery mechanisms. The 'additional checks' fix likely lives in Archive Utility or the TCC layer rather than in underlying POSIX attribute propagation paths — meaning new compression methods, nested archives, or non-standard metadata streams may bypass the mitigation without requiring a novel vulnerability. Audit your fleet for: (1) third-party extraction utilities that may not receive Apple's format-specific patches, (2) security tools that treat quarantined-but-unsigned code as lower-risk, and (3) MDM policies that rely on Gatekeeper as a hard stop rather than a defense-in-depth layer. Assume this bypass will recur with new archive formats — plan your detection and response around the pattern, not just this instance.

Reviewed through automated stages and approved by a human before publication.