In a latest security partnership with Mozilla, Anthropic discovered 22 separate vulnerabilities in Firefox — 14 of them categorised as “high-severity.” Many of the bugs have been mounted in Firefox 148 (the model launched this February), though just a few fixes should look ahead to the following launch.
Anthropic’s workforce used Claude Opus 4.6 over the span of two weeks, beginning within the javascript engine after which increasing to different parts of the codebase. In line with the publish, the workforce centered on Firefox as a result of “it’s each a posh codebase and one of the well-tested and safe open-source tasks on the planet.”
Notably, Claude Opus was a lot better at discovering vulnerabilities than writing software program to use them. The workforce ended up spending $4,000 in API credit attempting to concoct proof-of-concept exploits, however solely succeeded in two instances.
Nonetheless, it’s a reminder of how highly effective AI instruments will be for open-source tasks — even when they convey a flood of unhealthy merge requests alongside the helpful ones.
