Anthropic's coordinated vulnerability disclosure dashboard (red.anthropic.com)

Anthropic has launched a coordinated vulnerability disclosure dashboard to enhance security in open-source software, having already disclosed 1,596 vulnerabilities across 281 projects since February 2026. Using the Claude Mythos Preview, they partnered with external security firms to rigorously assess and report critical vulnerabilities under their disclosure policy. As of May 22, 2026, 97 vulnerabilities have been patched, with 88 assigned a Common Vulnerabilities and Exposure (CVE) record or GitHub Security Advisory, showcasing the successful identification and management of potential security threats. The significance of this initiative lies in its structured approach to security, utilizing AI for vulnerability detection while also ensuring human oversight in validation and reporting. This dual-layer process not only enhances the reliability of the disclosures but also aims to improve community trust in open-source software security practices. The dashboard will soon include filtering options by severity, providing developers and maintainers with clearer insights into vulnerabilities and the urgency of required fixes, ultimately fostering a more secure software ecosystem.