Claude, Codex, and other AI tools credited in today’s Apple security releases - 9to5Mac

What the report says

9to5Mac reported that Apple’s July 27, 2026 security documentation for a broad set of operating system updates shows growing use of AI systems in vulnerability research. The releases covered iOS 26.6, iPadOS 26.6, macOS Tahoe 26.6, macOS Sequoia 15.7.8, macOS Sonoma 14.8.8, tvOS 26.6, visionOS 26.6 and watchOS 26.6. After the software updates became available, Apple added technical notes describing affected components, possible security impact, fixes and researchers credited with reporting issues.

According to 9to5Mac, Apple’s credits include Anthropic researchers and Claude for issues tied to WebKit, WebKit Storage and WebDAV, in some cases alongside researchers from Calif.io. The article notes that Calif.io previously said in May that it used Anthropic’s Mythos Preview model to help produce a working macOS kernel memory-corruption exploit on M5 silicon within five days. Calif.io was also credited in multiple kernel fixes across Apple’s newly released updates.

The report said Apple listed 30 distinct CVEs across the day’s operating system releases for kernel vulnerabilities, with numerous researchers credited, suggesting overlapping reports of related flaws. 9to5Mac also noted that Apple has had access to Claude Mythos Preview through Anthropic’s Project Glasswing since April, while cautioning that public acknowledgments may not reflect every internally found issue.

Beyond Anthropic, Apple’s notes credited work involving OpenAI Codex Security, Z.AI’s GLM and NVIDIA’s AI Red Team, among others. The development matters because Apple’s public security acknowledgments indicate AI-assisted methods are becoming a visible part of how vulnerabilities are discovered and reported, potentially affecting patch planning and bug bounty programs.