Secure Agentics was named one of ten finalists in the techUK Cyber Innovation Den 2025, the UK's flagship pitching competition for emerging cyber security companies. Selected through a national open call, we pitched Adrian to a panel of industry judges on 15 October 2025.

Adrian is our open-source runtime security tool for AI agents. It analyses an agent's activity and reasoning to detect and block prompt injection, malicious tool use, and out-of-remit behaviour at runtime, helping teams adopt agentic AI safely.

About Secure Agentics. Secure Agentics builds Adrian, the open-source runtime security tool for AI agents, helping teams adopt agentic AI safely.