Introduction¶
AI-augmented security operations, built on the principle that human judgment and machine precision outperform either alone.
Centaur Security Labs is the research and development home of ARCHER - a local-first AI security operations agent designed for professional and enterprise environments.
The name comes from a model Garry Kasparov developed after his match against Deep Blue - human judgment and machine precision working together, each handling what it does best. The full story is here. The short version: that model is how we think about AI in security operations.
What is the project?¶
ARCHER is an agent that runs on local hardware, operates within your security boundary, and covers the full scope of security operations work - penetration testing, threat hunting, digital forensics, malware analysis, system hardening, and more - each domain built to the governing professional standards, each validated with real operational runs.
It is not a chatbot wrapper. It is not a prompt-engineered demo. It is a structured system where AI handles the three things AI is genuinely good at - generating candidate actions, interpreting heterogeneous tool output, and reasoning across multi-turn investigative chains - and deterministic code handles everything else.
Where to Start¶
New here? Start with What is ARCHER? - the vision, the design principles, and the problem it solves.
Technical audience? Go to V1 to V2 Evolution - the hardware realities, the architectural evolution, and the skill domain system.
Security leadership? Go to Trust Requirements and NIS2 & DORA Alignment - what "production-ready" means for a regulated environment.
Following the build? The Build Journal documents the full development story: design decisions, hard lessons, and the path to production.
ARCHER is currently in active development. This site is private until the first public release. All nine domains will be complete, validated, and documented before that release.