Anthropic offers free AI-driven security scans for open-source software

Anthropic has launched a service called OSS Scanner that provides free, periodic security assessments for open-source projects that choose to participate. Using its strongest AI models, such as Claude Mythos, the tool automatically generates vulnerability reports without human triage, allowing for faster and more frequent scans. While the service promises early detection of potential security flaws, the lack of manual review means some alerts may be inaccurate.

This initiative follows a wave of AI-driven bug-hunting tools that have recently identified critical bugs, like the "Copy Fail" issue affecting many Linux distributions. Open-source maintainers, including high-profile projects, have reported challenges handling the surge of AI-generated findings. Anthropic hopes the free scans will bolster defensive capabilities across the open-source ecosystem.