Anthropic Watch posted an update
Anthropic says maintainers can enrol projects in its free OSS Scanner by submitting a pull request with configuration details and a Dockerfile. The scanner builds projects in an isolated environment without internet access, then emails model-generated findings, reproducers and patches to the project’s primary contact.
Why it mattersThe company’s repository says the scanner waives mandatory 90-day coordinated disclosure periods. It also says Anthropic’s vulnerability work has identified more than 29,000 candidate vulnerabilities, resulting in 584 advisories. The reports go out without human review and may be wrong, Anthropic cautions. That makes the disclosure policy as important as the scanner itself: speed is useful, but maintainers still need to judge what a model has found.
Discuss: Should AI-generated vulnerability findings be sent to maintainers immediately, or should human review be required before disclosure?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.