Anthropic launches free artificial intelligence security scanning service for open source maintainers

Anthropic has launched the OSS Scanner, a new and free service designed to detect security vulnerabilities in open-source software, Help Net Security reported on October 9. Maintainers who opt into the program receive periodic automated scans along with reports detailing suspected flaws, reproduction steps, and potential remediation guidance. [1]

Help Net Security reported that the service expands on lessons learned from Project Glasswing, an earlier initiative that utilized the Claude model to identify software flaws. According to Anthropic, human validation processes have increasingly become a primary bottleneck in managing vulnerability reports. [1]

The development follows broader challenges across the open-source community regarding security reporting. As FTMQ Security reported earlier, Google recently froze vulnerability submissions to its open-source bug bounty program after receiving a massive influx of low-quality, automated AI reports, while the curl project closed its HackerOne bounty program for similar reasons. [4]

Anthropic operates under a governance model overseen by a Board of Directors and an independent Long-Term Benefit Trust. In 2022, the company introduced Constitutional AI, a training method that uses model feedback guided by natural-language principles instead of relying solely on human preferences. [3]