Irregular: Frontier Security Research and AI Evaluation
Irregular is a frontier-security lab building research platforms for real-world AI security scenarios. It was Anthropic's third-party cybersecurity-evaluation partner in the July 2026 incidents.
kindResearch lab
siteirregular.com
Sources
Irregular
Docs
About Irregular
We build next-generation defenses through high-fidelity research platforms that simulate and monitor real-world AI security scenarios.
Anthropic
Announcement · Jul 30, 2026
Investigating three real-world incidents in our cybersecurity evaluations
the evaluation environment of Irregular, one of our third-party evaluation partners