Responsible Disclosure

On 7 April 2026, Anthropic announced two things at once. The first was a new frontier model called Claude Mythos Preview. The second was Project Glasswing, a coalition of twelve technology and finance companies that would receive controlled access to that model, with everyone else, including paying API customers, locked out indefinitely. The accompanying blog posts from Anthropic’s red team made a remarkable claim: Mythos Preview, given an isolated container and a vague prompt, had autonomously discovered thousands of zero-day vulnerabilities across “every major operating system and every major web browser”: a 27-year-old denial-of-service bug in OpenBSD’s TCP stack, a 17-year-old remote code execution flaw in FreeBSD’s NFS server, fully weaponised end-to-end. A 16-year-old vulnerability in FFmpeg’s H.264 codec that had survived every fuzzer and every human reviewer to look at the code since 2010. In Mozilla Firefox alone, the model surfaced 271 zero-days, shipped as fixes in Firefox 150, the largest single batch of security fixes in the browser’s history. ...

In the spring of 2021, Dutch Institute for Vulnerability Disclosure (DIVD) researcher Wietse Boonstra quietly uncovered seven critical flaws in Kaseya’s widely used IT management software. DIVD warned the company within days, flagging more than 2,200 vulnerable systems across the globe. Weeks later, three flaws remained unpatched, and the REvil ransomware gang pounced. Overnight, some 1,500 organisations were paralysed, from supermarkets in Sweden to schools in New Zealand. This was not an isolated close call. In a 2023 study with the University of Twente, DIVD found that less than half of Dutch municipalities acted promptly when notified of exploitable flaws in their email systems. In some cases, local authorities ignored the warnings entirely. ...