Claude's Firefox Audit Exposes Hundreds of Missed Security Flaws

Anthropic's Claude AI found 271 vulnerabilities in Firefox during a two-week audit, including 14 serious security flaws—one-fifth of all critical bugs Mozilla patched in 2025. Analysis: The sheer number and speed of discovery suggest that traditional human security reviews, even on widely-used open-source software, may no longer be enough. The finding raises a hard question for organizations: if mature, heavily-reviewed code still conceals hundreds of flaws, what gaps exist in your own security practices.