Firefox engineers have fixed 22 security vulnerabilities after Anthropic's Frontier Red Team used AI to scan the browser's codebase. The collaboration revealed 14 high-severity bugs and 90 additional issues across Firefox's JavaScript engine and broader codebase. All vulnerabilities have been patched in the recently released Firefox 148.

This marks a significant milestone in AI-assisted security testing, as Anthropic's team used Claude to identify bugs that traditional methods missed. The AI-generated reports included minimal test cases that allowed Mozilla's security team to quickly verify and reproduce each issue. Within hours of receiving the findings, engineers began landing fixes.

The scale of discoveries is particularly notable given Firefox's reputation as one of the most scrutinized codebases on the web. Despite decades of fuzzing, static analysis, and regular security reviews, the AI model uncovered numerous previously unknown bugs. Anthropic has published a technical write-up of their research process, demonstrating how large-scale AI analysis can complement existing security tools.