Anthropic’s Claude AI Uncovers 22 Vulnerabilities in Firefox in Just Two Weeks
In a recent collaboration between Anthropic and Mozilla, the AI model Claude Opus 4.6 identified 22 vulnerabilities within the Firefox browser over a two-week period. Of these, 14 were deemed high-severity. Most of these issues have been addressed in Firefox version 148, released in February, with the remaining fixes scheduled for upcoming updates.
The initiative began with Claude Opus 4.6 analyzing Firefox’s JavaScript engine before expanding to other parts of the codebase. Mozilla’s choice to focus on Firefox was strategic, given its complexity and reputation as a secure open-source project.
While Claude demonstrated proficiency in identifying vulnerabilities, it faced challenges in creating exploitative software. The team invested $4,000 in API credits to develop proof-of-concept exploits but succeeded in only two instances.
This collaboration underscores the potential of AI tools in enhancing open-source projects, despite the occasional influx of less useful contributions.
