Provable Coordination for LLM Agents via Message Sequence Charts

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bollig, Benedikt, Függer, Matthias, Nowak, Thomas
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914516078428160
author Bollig, Benedikt
Függer, Matthias
Nowak, Thomas
author_facet Bollig, Benedikt
Függer, Matthias
Nowak, Thomas
contents Multi-agent systems built on large language models (LLMs) are difficult to reason about. Coordination errors such as deadlocks or type-mismatched messages are often hard to detect through testing. We introduce a domain-specific language for specifying agent coordination based on message sequence charts (MSCs). The language separates message-passing structure from LLM actions, whose outputs remain unpredictable. We define the syntax and semantics of the language and present a syntax-directed projection that generates deadlock-free local agent programs from global coordination specifications. We illustrate the approach with a diagnosis consensus protocol and show how coordination properties can be established independently of LLM nondeterminism. We also describe a runtime planning extension in which an LLM dynamically generates a coordination workflow for which the same structural guarantees apply. An open-source Python implementation of our framework is available as ZipperGen.
format Preprint
id arxiv_https___arxiv_org_abs_2604_17612
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Provable Coordination for LLM Agents via Message Sequence Charts
Bollig, Benedikt
Függer, Matthias
Nowak, Thomas
Programming Languages
Artificial Intelligence
Multi-agent systems built on large language models (LLMs) are difficult to reason about. Coordination errors such as deadlocks or type-mismatched messages are often hard to detect through testing. We introduce a domain-specific language for specifying agent coordination based on message sequence charts (MSCs). The language separates message-passing structure from LLM actions, whose outputs remain unpredictable. We define the syntax and semantics of the language and present a syntax-directed projection that generates deadlock-free local agent programs from global coordination specifications. We illustrate the approach with a diagnosis consensus protocol and show how coordination properties can be established independently of LLM nondeterminism. We also describe a runtime planning extension in which an LLM dynamically generates a coordination workflow for which the same structural guarantees apply. An open-source Python implementation of our framework is available as ZipperGen.
title Provable Coordination for LLM Agents via Message Sequence Charts
topic Programming Languages
Artificial Intelligence
url https://arxiv.org/abs/2604.17612