Anthropic Launches Free AI Security Scanning Service for Open-Source Projects
Anthropic’s OSS Scanner will run free, periodic vulnerability scans with models including Claude Mythos, giving resource-strapped projects earlier warnings. Reports receive no human review or triage and may be invalid, adding to AI-generated submissions that have already strained open-source maintainers.
Anthropic announced in October 2026 a free security scanning service called OSS Scanner, designed to help open-source projects identify vulnerabilities using the company’s AI models.
Open-source projects that opt in will receive “thorough, periodic security scans” run by Anthropic’s strongest models, including Claude Mythos, at no cost. The scans are intended to alert maintainers to potential security issues faster and more frequently than traditional methods might allow.
There is a notable trade-off, however. Anthropic has been explicit that OSS Scanner’s reports are fully model-generated and will not undergo human review or triage. The company acknowledged this means reports could be incorrect or invalid. “This will enable faster and more frequent scanning,” Anthropic stated, “but means that it is possible reports will be incorrect or invalid.”
OSS Scanner enters a space where AI-assisted bug hunting has already produced results. AI tools have helped uncover significant security flaws in open-source software in recent months, including the “Copy Fail” bug, which affected nearly every Linux distribution in May 2026.
At the same time, the rise of AI-generated bug reports has created challenges for some open-source maintainers. Projects associated with Linus Torvalds and even Google have reportedly struggled to manage the volume of AI-generated submissions.
For open-source projects with limited resources, OSS Scanner could provide earlier warnings about security weaknesses. The absence of human review, however, may require maintainers to invest their own time in validating any flagged issues before acting on them.