Imperum has been approved for Anthropic’s Cyber Verification Program (CVP). The company’s announcement identifies vulnerability validation, detection engineering, threat hunting, incident response and forensic investigation as defensive work that can involve dual-use activities.
What the approval supports
Anthropic describes CVP as a way for qualifying security professionals to conduct legitimate cybersecurity work with certain cyber-misuse safeguards adjusted. The approval supports Imperum’s work on governed AI for security operations.
In its announcement, Imperum also describes testing defensive AI agents against simulated adversarial scenarios, reviewing the outcomes and repeating the tests to improve their performance.