Inside the race to adapt to an AI-powered security world

XBOW’s AI-driven offensive security platform uncovered multiple real issues in Moderna’s development environment — a WAF bypass, a hard-coded API key in source code, and an API susceptible to SQL injection that cascaded into an outage for dependent services — demonstrating how advanced AI models can rapidly discover exploitable flaws. The report frames this as part of a broader industry shift: more capable models (e.g., Claude Mythos) are surfacing large volumes of vulnerabilities faster than organizations can remediate them, driving adoption of mitigations like eBPF-based compensating controls.