NARVI

About

Research

Projects

Journal Notes News CV

Research Paper Closed-world tool grounding Can an agentic system be made structurally incapable of claiming a capability it doesn't have? Case study: A.R.C.H.O.N.'s closed tool manifest.
Project A.R.C.H.O.N. Continued development — multi-agent orchestration, self-critique gating, evaluation harness.
Project PRISM Continued development — organ classification system, permission-gated data layer.
Note Tying it together How the closed-world grounding question shows up concretely in both A.R.C.H.O.N. and PRISM.

Notes

Short dispatches on applied AI and tech, as I run into them.

Nothing here yet.

Applied AI, agent orchestration, and notes from building A.R.C.H.O.N. and PRISM.

N.A.R.V.I. — Nexus Applied Research & Verified Intelligence

Site founded September 15, 2026