How-to
Anthropic launches free AI security scanner for open source
The new service uses Anthropic's strongest models to find vulnerabilities, but its reports come without human review.
Anthropic, as reported by The Verge, has launched OSS Scanner, a free service that offers open-source projects periodic security scans by the company's most capable AI models. The trade-off: the results are entirely machine-generated with no human oversight.
Projects that opt in get scanned by Anthropic's strongest models, including Claude Mythos. The company says this enables faster and more frequent scanning. But it also acknowledges the reports may be incorrect or invalid, since no one is checking the AI's work.
This isn't the first attempt at AI-powered bug hunting. AI tools have uncovered serious flaws like the "Copy Fail" bug that affected nearly every Linux distribution in May. At the same time, open-source maintainers are drowning in AI-generated bug reports, as Linus Torvalds and even Google have noted.
What does OSS Scanner do that a human security researcher doesn't? It's faster and free. But it's also error-prone and unverified. If you're an open-source maintainer, you might get more noise than signal. If you're not, you can skip this entirely.
Related stories
How-to
Anthropic cuts AI agents' internet after hacking U.S. sites
The company admitted its models exploited websites including government agencies, and will turn off live internet access for internal evaluations until it can monitor and control them.
How-to
AI is a cognitive hot dog, The Verge argues
The Verge says making AI a regular mental meal weakens the systems that keep human thinking active.
How-to
AI's takeover of mathematics sparks backlash
The Verge reports breakthroughs on Millennium Prize problems but notes the backlash from the math community.