Skip to content
#explainable ai Open access

FCoP: A File-Based Coordination Protocol for Governed Multi-Agent Systems

Sep 2026 · Zenodo (CERN European Organization for Nuclear Research)

Abstract

Multi-agent work often outlives individual sessions and processes. A successor must identify not only an output but also the current task and attempt, the results submitted and accepted, and the evidence and authority for the next change. This paper systematically describes and studies FCoP 4.0, a protocol that expresses these coordination obligations in a project-visible filesystem workspace. Stable identities, four business-document types, path-authoritative lifecycle state, attempts, evidence binding, authorization, branch convergence, and durable receipts form an executable work contract that separates formal coordination records from transient actors. We explain the composition rationale and its tradeoffs, analyze freshness of completed commits under explicit storage, writer, and authority assumptions, and distinguish current-content coverage from acceptance-decision coverage. Using the frozen reference implementation, an end-to-end handoff proceeds through five successively terminated processes in each of three fresh workspaces, completing rejection, revision, acceptance, convergence, and archival. Four controlled cross-process commit orders, repeated three times each, exhibit the expected outcomes. Small-workspace measurements characterize local synthetic execution costs. Together with retained historical tests, failures, and boundary findings, the results support FCoP as an implementable, inspectable, and transferable persistent coordination protocol. They do not equate protocol acceptance with business truth or extend local validation to arbitrary Hosts or distributed deployments. This record contains the English paper, its Chinese edition, editable manuscript sources, and the frozen research supplement for one study. Preprint version 1.0; not peer reviewed. The earlier TMPA Architecture Paper A1.0, published in the TMPA V1.0 Publication System (DOI: 10.5281/zenodo.21888488), already introduced the architecture context and a file-based FCoP reference instantiation. This paper focuses on the concrete FCoP 4.0 contract and the frozen 4.0.3 implementation. It acknowledges that predecessor without claiming those foundational ideas as first introduced here or treating older TMPA/CodeFlowMu results as new-version conformance evidence. The supplement preserves original test failures, controlled-schedule interruptions, API findings, and environment-blocked installation preparation. Different evidence layers are not pooled into a single pass count. Experiments used synthetic workloads, one frozen implementation, and a local test Profile; they do not establish independent certification, live-Host interoperability, business truth, or arbitrary distributed deployment guarantees. AI assistance in research and manuscript preparation is disclosed in the paper. © 2026 Wei Zhu. The paper, original figures, and newly authored research materials are licensed under Creative Commons Attribution 4.0 International (CC BY 4.0): https://creativecommons.org/licenses/by/4.0/. Copying, sharing, adaptation, and commercial reuse are permitted with appropriate attribution, a license link, and an indication of changes; no endorsement may be implied. The frozen FCoP source remains under its original MIT license. Third-party material, quoted extracts, and other separately licensed material retain their own terms. This grant covers only rights the author can license and places no restrictions on public-domain facts.

View source

Similar papers

#artificial intelligence Conference Open access Apr 2020

ECCOLA - a Method for Implementing Ethically Aligned AI Systems

The method, ECCOLA, is presented, which aims at making the high-level AI ethics principles more practical, making it possible for developers to more easily implement them in practice.

Ville Vakkuri, Kai-Kristian Kemell, P. Abrahamsson · 64 citations · ⚡6
#computer vision Review Apr 2024

AI-powered Code Review with LLMs: Early Results

The goal is to not only refine the accuracy of the LLM-based tool but also to underscore its potential in streamlining the software development lifecycle through proactive code improvement and education.

Z. Rasheed, Malik Abdul Sami, Muhammad Waseem et al. · 62 citations · ⚡3
#computer vision Open access Mar 2024

LLM-based agents for automating the enhancement of user story quality: An early report

The use of large language models to automatically improve the user story quality in Austrian Post Group IT agile teams is explored, with a reference model for an Autonomous LLM-based Agent System developed and implemented at the company.

Zheying Zhang, M. Rayhan, Tomas Herda et al. · 48 citations · ⚡4
#computer vision Review Mar 2024

System for systematic literature review using multiple AI agents: Concept and an empirical evaluation

This paper introduces a novel multi-AI-agent system designed to fully automate SLRs, and demonstrates how it substantially reduces the time and effort traditionally required for SLRs while maintaining comprehensiveness and precision.

Abdul Malik Sami, Z. Rasheed, Kai-Kristian Kemell et al. · 44 citations · ⚡2
#computer vision Feb 2024

Can Large Language Models Serve as Data Analysts? A Multi-Agent Assisted Approach for Qualitative Data Analysis

The proposed LLM-based multi-agent system automates qualitative data analysis process, creating opportunities for researchers and practitioners, and future improvements focus on enhancing multilingual performance and integrating continuous expert feedback.

Z. Rasheed, Muhammad Waseem, Aakash Ahmad et al. · 41 citations
#artificial intelligence Conference Open access Jun 2018

The Key Concepts of Ethics of Artificial Intelligence

It is suggested that the focus on finding keywords is the first step in guiding and providing direction for future research in the AI ethics field.

Ville Vakkuri, P. Abrahamsson · 39 citations · ⚡2

Related blog posts

Microsoft Research Blog Oct 7, 2026

Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses

Training AI agents with reinforcement learning can be challenging because their tools, context, and decision-making are managed by complex frameworks. Agent Lightning connects existing agents to RL training, making it easier to improve them without rebuilding them. The post Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.