How Anthropic Claude Mythos is reshaping the vulnerability landscape

Anthropic's Claude Mythos Preview, unveiled in April 2026, is a frontier AI model capable of autonomously discovering and chaining zero-day vulnerabilities across major OSes and browsers, compressing the window from discovery to weaponization from weeks to hours. Traditional static scanners and pipeline tools cannot keep pace because they lack production context and generate unactionable alert volumes. Runtime security approaches — continuously monitoring which libraries are actually loaded, performing reachability analysis on execution paths, and mapping business impact via live topology graphs — are positioned as the necessary response. Dynatrace promotes its OneAgent, Smartscape topology graph, Grail data lakehouse, and proprietary vulnerability feed as the platform to filter exploitable, business-critical CVEs from the noise, especially as open-source attack surfaces grow fastest under AI-driven discovery.