Anthropic Says Mythos Found 23,000 Potential Vulnerabilities in Open-Source Projects
Summary
Anthropic says Mythos Preview found 23,000+ potential vulnerabilities across 1,000+ open-source projects. External firms confirmed 1,726, including 1,000+ high- or critical-severity issues; 75 high- or critical-severity issues have been patched.
Key points
- Anthropic says Mythos Preview identified more than 23,000 potential vulnerabilities across over 1,000 open-source projects.
- External security firms reviewed 1,900 findings and confirmed 1,726, including more than 1,000 rated high or critical.
- Anthropic estimates nearly 3,900 high- and critical-severity findings may be confirmed from current results; it says ongoing scans could raise the total number of severe findings to 6,200.
- Anthropic says more than 1,100 unverified findings were reported to vendors; 75 high- or critical-severity issues have been patched, and vendors published 65 advisories.
- Around 50 organizations have Mythos Preview access through Project Glasswing. Mozilla reported finding 271 Firefox vulnerabilities; Google’s recent Chrome detections could involve Mythos, Google’s own AI tools, or both.
- Anthropic says it has not yet developed safeguards strong enough to prevent Mythos misuse.
Article Details
- Event Type
- Anthropic-reported AI-assisted vulnerability discovery across more than 1,000 open-source software projects.
- Impact
- Anthropic says Mythos Preview identified more than 23,000 potential vulnerabilities. Of 1,900 externally reviewed findings, 1,726 were confirmed, including more than 1,000 rated high or critical. Anthropic estimates nearly 3,900 high- and critical-severity vulnerabilities may be confirmed from current findings, with the total potentially reaching 6,200 as scanning continues. More than 1,100 unverified findings were reported to vendors; 75 critical- or high-severity issues had been patched, and vendors had published 65 advisories. The findings add pressure to an already overloaded security ecosystem, while Anthropic says safeguards against misuse are not yet strong enough.
Vendors
AnthropicAnthropic says its Claude Mythos model discovered thousands of severe vulnerabilities across more than 1,000 open source software (OSS) projects.GoogleGoogle has also been given access, but it’s unclear whether the recent surge in Chrome vulnerability detections is due to Mythos, the company’s own AI tools, or both.MozillaMozilla reported finding 271 Firefox vulnerabilities, and Mythos has helped Palo Alto Networks find dozens of flaws.Palo Alto NetworksMozilla reported finding 271 Firefox vulnerabilities, and Mythos has helped Palo Alto Networks find dozens of flaws.
Products
Claude Mythos PreviewAccording to the AI giant, Mythos Preview has identified more than 23,000 potential vulnerabilities.Claude SecurityIn response to the surge in AI-powered vulnerability discovery, Anthropic recently unveiled Claude Security, a codebase scanner designed to help developers find security issues in their applications.curlMythos found only one low-severity vulnerability in Curl, with experts debating whether that is a failure of the AI model or a testament to the open source data transfer tool’s maturity.Mozilla FirefoxMozilla reported finding 271 Firefox vulnerabilities, and Mythos has helped Palo Alto Networks find dozens of flaws.