Secure AI for Real-World Systems

ASB Security is an open initiative focused on making LLMs and multi-agent systems secure in production.

ASB Security is an open initiative focused on making LLMs and multi-agent systems secure in production.

Join the Community

ASB Security is an open, collaborative project.

We welcome contributors from research, engineering, and industry who care about making AI safer in practice.

Researchers – to design evaluations, benchmarks, and attack/defense methodologies.
Engineers – to build and refine the gateway, rules engine, integrations, and dev tooling.
Organizations – to run pilots in real environments, provide feedback, and share requirements from production.