Permission before action.
Make authority explicit: who can act, within which scope, and under what conditions.
Building a verifiable foundation for what automated systems are allowed to do. And when they must stop.
Start a conversation ↗Research & development · United States
Our approach
We’re developing consent and authorization infrastructure for systems that act on behalf of people. Three principles guide the work.
Make authority explicit: who can act, within which scope, and under what conditions.
Treat revocation and STOP as foundational requirements, from the outset.
Design for portable receipts that make authorization decisions independently reviewable.
Current work
We develop research authorization software and evidence infrastructure. Our focus is how authority is granted, constrained, revoked, and recorded.
Our research examines authorization failures, delegation and revocation faults, stale state, retries, evidence integrity, and differences between formal models and runtime behavior.
Security testing is scoped to systems we own and operate, in non-public, access-controlled research environments.
Our research uses Lean 4, Isabelle/HOL, and SPARK/Ada to machine-check selected properties of the authorization and evidence design.
Selected artifacts can be made available to qualified reviewers under appropriate confidentiality. These results do not establish end-to-end verification of a deployed system.
Research & development. No public Nexus Spine product is currently offered through this site. The research systems described here are intended for non-public research, not customer-facing production use.
Let’s talk
For research, technical collaboration,
or partnership inquiries.