mirror of
https://github.com/lux-apps/papers.git
synced 2026-11-03 05:07:46 +00:00
zeekay
f4ee7a0525
Bylines, paper authorship, correspondence addresses and the founder blurb under each post named an individual. They name the organisation that published the work instead: the brand on posts and pages, the research byline on papers, and research@ in place of a personal address. Both are what these repositories already used elsewhere — nothing new was coined. Co-authors are untouched. Where a paper named someone else alongside, they stay named; only the one byline changed. Bibliography entries, running heads, PDF metadata and affiliations move with it, so a citation still resolves to whoever the cited paper names. Sample and fixture data uses the placeholder name these tests already use elsewhere, so no assertion changed meaning.
79 lines
3.9 KiB
TeX
79 lines
3.9 KiB
TeX
|
|
%% ===================================================================
|
|
%% LP-134 canonical naming (2025-12-15):
|
|
%% Per LP-134 (Lux Chain Topology), M-Chain has been split into:
|
|
%% M-Chain — MPC ceremonies for bridge custody of EXTERNAL wallets
|
|
%% (CGGMP21, FROST, Pulsar-general)
|
|
%% F-Chain — FHE compute + TFHE bootstrap-key generation (encrypted EVM)
|
|
%% M-Chain is REMOVED entirely; teleport is B-Chain (bridgevm, LP-6000). No teleportvm.
|
|
%% Where this paper says "M-Chain MPC" / "M-Chain threshold governance" /
|
|
%% "M-Chain TFHE" / "M-Chain custody", read as M-Chain (MPC) or F-Chain (FHE).
|
|
%% ===================================================================
|
|
|
|
\documentclass[11pt,a4paper]{article}
|
|
\usepackage{../shared/luxcover}
|
|
\usepackage[utf8]{inputenc}
|
|
\usepackage[T1]{fontenc}
|
|
\usepackage{amsmath,amssymb,amsfonts,amsthm}
|
|
\usepackage{graphicx}
|
|
\usepackage{hyperref}
|
|
\usepackage{algorithm}
|
|
\usepackage{algpseudocode}
|
|
\usepackage{booktabs}
|
|
\usepackage{multirow}
|
|
\usepackage{array}
|
|
\usepackage{xcolor}
|
|
\usepackage{float}
|
|
\usepackage{enumitem}
|
|
\usepackage{mathtools}
|
|
\usepackage{listings}
|
|
|
|
\newtheorem{theorem}{Theorem}
|
|
\newtheorem{lemma}[theorem]{Lemma}
|
|
\newtheorem{proposition}[theorem]{Proposition}
|
|
\newtheorem{definition}{Definition}
|
|
\newtheorem{corollary}[theorem]{Corollary}
|
|
\newtheorem{remark}{Remark}
|
|
|
|
\hypersetup{
|
|
colorlinks=true,
|
|
linkcolor=blue,
|
|
citecolor=blue,
|
|
urlcolor=blue
|
|
}
|
|
|
|
\title{\textbf{Agent Sovereignty: AI Agents as First-Class Sovereign Processes on Lux Network}\\[6pt]
|
|
\large Vaults, Identities, Capabilities, Budgets, and Private Memory\\for Autonomous AI on a Multi-Chain Trust Kernel}
|
|
|
|
\author{
|
|
Lux Research\thanks{Corresponding author: [email protected]}\\
|
|
\textit{Lux Industries}\\
|
|
\texttt{[email protected]}
|
|
}
|
|
|
|
\date{2025}
|
|
|
|
\begin{document}
|
|
|
|
\luxcoverpage
|
|
|
|
\begin{abstract}
|
|
We present a framework for treating AI agents as sovereign processes---first-class entities on Lux Network with their own encrypted vaults, decentralized identities, fine-grained capabilities, economic budgets, and private persistent memory. Existing agent frameworks (AutoGPT, LangChain, CrewAI) treat agents as stateless API wrappers: ephemeral processes that hold no assets, own no identity, and leave no verifiable trace. This is fundamentally inadequate for agents that must persist across sessions, manage funds, sign transactions, coordinate with other agents, and be held accountable for their actions. We introduce the \emph{Agent Capsule}---a self-contained sovereignty boundary comprising an encrypted SQLite vault, a DID identity anchored on the Lux I-Chain, scoped capability tokens, threshold spend controls enforced by the M-Chain, and CRDT-synchronized private memory with optional FHE encryption. We formalize the capability delegation model, prove containment guarantees under agent compromise, define the economic settlement protocol for per-action cost accounting, and specify the private inference pipeline using CKKS-encrypted model inputs with threshold decryption of results. The architecture composes with the Lux trust kernel (P-Chain for validator sets, C-Chain for smart contracts, M-Chain for threshold cryptography, I-Chain for identity) and integrates with the Model Context Protocol (MCP) for tool use. Agent Capsules transform AI agents from disposable scripts into persistent, auditable, economically rational sovereign processes---a Web5 primitive, not a chatbot wrapper.
|
|
\end{abstract}
|
|
|
|
\input{sections/01-introduction}
|
|
\input{sections/02-agent-capsule}
|
|
\input{sections/03-agent-identity}
|
|
\input{sections/04-agent-memory}
|
|
\input{sections/05-agent-capabilities}
|
|
\input{sections/06-agent-economics}
|
|
\input{sections/07-private-inference}
|
|
\input{sections/08-multi-agent-coordination}
|
|
\input{sections/09-agent-attestation}
|
|
\input{sections/10-security-model}
|
|
\input{sections/11-implementation}
|
|
\input{sections/12-related-work}
|
|
\input{sections/99-conclusion}
|
|
|
|
\end{document}
|