Anthropic rolls out free AI security scans for open source

Artificial intelligence startup Anthropic launched a security auditing service called OSS Scanner. The tool provides automated vulnerability scans at no cost to open-source software projects that choose to participate.

The company stated that the vulnerability reports will be generated entirely by its most capable artificial intelligence models, including Claude Mythos. Anthropic noted that the outputs will not undergo human review or triage before reaching developers. This design allows for faster and more frequent scanning cycles, though the company warned that some findings may turn out to be incorrect or invalid.

AI systems have previously uncovered critical flaws in community-maintained software, such as the Copy Fail vulnerability that affected nearly every Linux distribution in May. However, the surge in automated vulnerability reporting has created backlogs for developers, prompting concerns from project leaders including Linus Torvalds and technology firms like Google.