CVE-2026-43745
The 'improved input validation' fix language for CVE-2026-43745's out-of-bounds write should concern you more than the 6.5 CVSS score suggests. When Apple describes a memory corruption fix as 'added bounds checking,' they're indicating structural changes to memory management. 'Improved input validation' means the code was attempting to validate its way to memory safety—a design pattern that has a documented failure mode: the bypass. An out-of-bounds write in a parsing layer that requires input validation to prevent is categorically different from a classic buffer overflow, and it signals that malformed input was reaching an unsafe write operation before the patch. That doesn't mean exploitation is trivial; it means exploitation requires specific, non-generic payloads that pattern-match against the validation logic. The crash-only impact statement is equally worth scrutinizing—it likely reflects what Apple confirmed internally, not what's achievable. Apple discovered this internally (the simultaneous seven-OS release is a strong signal of internal discovery versus external reporting), which means they fuzzed for crashes without the adversarial pressure of a proof-of-concept forcing them to prove code execution. For an out-of-bounds write in WebKit—a component that has delivered far worse than crashes historically—the conservative disclosure is notable but not reassuring. Your priority actions: First, map your WebKit exposure beyond Safari. This affects every WKWebView instance in your iOS, macOS, and visionOS deployments, including third-party browsers (legally mandated to use WebKit on iOS), Electron apps, and embedded web views in enterprise applications. Second, watch for follow-on disclosures. Historical patterns show validation fixes in WebKit rendering paths cluster within 90-120 days—Apple's systematic code review likely triggered broader auditing. Third, account for version fragmentation in your patching logic. The 26.5.2 (iOS/macOS) versus 26.6 (tvOS/watchOS/visionOS) split means some devices are provably behind, and attackers reverse-engineering the patch can immediately identify which targets are exposed. Fourth, recognize that enterprise deployment chains extend the exposure window well past the CVE publication date—custom apps with embedded WebKit views don't update via App Store and may remain vulnerable while you wait for enterprise-signed deployments to propagate.
Reviewed through automated stages and approved by a human before publication.