Metraipolis › archive › dc_4eab34c6f89e
Where Does Exactly-Once Live? Model, Harness, and Tool-Contract Effects on Duplicate Side Effects in LLM Agents — arXiv:2609.29095
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.29095
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T04:27:17.362Z
- SHA-256 of the archived text
c513af3b16820c6d758a61f9d55ad6bcf67a861e1a5df6948eb4239e587147e5- Stable URL of this page
https://metraipolis.com/documento/dc_4eab34c6f89e
How to cite
Where Does Exactly-Once Live? Model, Harness, and Tool-Contract Effects on Duplicate Side Effects in LLM Agents — arXiv:2609.29095 — arXiv (en), https://arxiv.org/abs/2609.29095 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4eab34c6f89e (Metraipolis archive dc_4eab34c6f89e, sha256 c513af3b16820c6d758a61f9d55ad6bcf67a861e1a5df6948eb4239e587147e5)
BibTeX
@misc{dc4eab34c6f89e,
title = {Where Does Exactly-Once Live? Model, Harness, and Tool-Contract Effects on Duplicate Side Effects in LLM Agents — arXiv:2609.29095},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.29095},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4eab34c6f89e; sha256 c513af3b16820c6d758a61f9d55ad6bcf67a861e1a5df6948eb4239e587147e5},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.29095. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
When a tool-using agent's write times out or returns a server error, the action may already have taken effect. Retrying blindly duplicates it -- a second charge, a second announcement, a second deployment -- while giving up skips required work. We ask where exactly-once behaviour should be enforced: in the model, in the agent harness, or in the tool contract. We introduce LIMBO, a deterministic sandbox of six services with realistic contracts (optional idempotency keys, eventually consistent and missing read paths) and twelve fault modes injected at the service boundary, including late commits, redelivery and partial batches; every episode is graded against a ledger of committed effects. Across 25,930 episodes spanning nine recent models, three production agent harnesses, two contract variants and fifteen recovery conditions, the answer depends on the fault. When an immediate read-back can reveal what happened, the model decides: frontier models instructed to act exactly once almost never duplicate a write whose acknowledgement was lost (0.5%), weaker models often do, and the model explains 53% of the explained variance. When it cannot -- the request is still in flight, or the transport delivered it twice -- the same frontier models duplicate in 56% and 74% of episodes, and the contract explains 81%. We prove that no verification-only policy is exactly-once under late commits without a bound on in-flight time. Waiting works when such a bound is short and known, but with heavy-tailed in-flight delays even an hour of waiting per episode falls short of offering an idempotency key on every write, which lowers the duplicate rate from 28% to 4% because agents use keys when they exist. The harness barely matters, a guard that attaches keys transfers across harnesses unchanged, and agents reported success in 90% of the episodes in which they had duplicated an effect. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-25 in cs.AI). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2609.29095 — 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 2026-10-08T03:50:58.005Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).