Metraipolis › archive › dc_6c8b1bb8bcb8

Concordat: Typed Sessions for LLM Agents — arXiv:2610.37461

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.37461
License (as archived)
CC0-1.0
Archived
2026-10-28T03:02:07.422Z
SHA-256 of the archived text
d579a9303c0a16a862517c67895ae5b46303f127798aa0feb85ac90a184c1798
Stable URL of this page
https://metraipolis.com/documento/dc_6c8b1bb8bcb8

How to cite

Concordat: Typed Sessions for LLM Agents — arXiv:2610.37461 — arXiv (en), https://arxiv.org/abs/2610.37461 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6c8b1bb8bcb8 (Metraipolis archive dc_6c8b1bb8bcb8, sha256 d579a9303c0a16a862517c67895ae5b46303f127798aa0feb85ac90a184c1798)
BibTeX
@misc{dc6c8b1bb8bcb8,
  title = {Concordat: Typed Sessions for LLM Agents — arXiv:2610.37461},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.37461},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6c8b1bb8bcb8; sha256 d579a9303c0a16a862517c67895ae5b46303f127798aa0feb85ac90a184c1798},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.37461. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Agents built on large language models are increasingly allowed to take consequential actions: cancelling bookings, issuing refunds, changing accounts. The rules that govern those actions, such as "obtain the customer's explicit confirmation before changing anything", are usually stated in the prompt and followed only as reliably as the model chooses to follow them. We present Concordat, a Python library that treats an agent as an object whose interface is a finite-state protocol, in the spirit of typestate and session types, and enforces that protocol on the model at run time. On every turn the model is offered only the tools that are legal in the current protocol state, calls outside that set are rejected before they have side effects, decisions reserved for humans can only be requested, and token and tool-call budgets are charged and split linearly across handoffs. Protocols are checked when they are defined, and a subclass of an agent is rejected if it could bypass its parent's protocol, a behavioural form of Liskov substitution. We evaluate a Concordat agent against the standard agent of tau-bench on its 50 airline tasks, adding simulated customers who pressure the agent to skip confirmation. The standard agent made changes without the required "yes" in 13 of 50 tasks with gpt-4.1 and in 29 of 128 task-runs with Claude Haiku 5.5; the Concordat agent made none, with no significant difference in task success. tau-bench's own reward did not detect most of these violations. Concordat is open source.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.MA, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2610.37461 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-28T03:01:03.539Z dalla pagina https://arxiv.org/list/cs.MA/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).