Metraipolis › archive › dc_2affcaec0e7c

Vestrum: Improving Agent Harnesses by Adapting Their Verification, Structure and Memory — arXiv:2609.33822

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.33822
License (as archived): CC0-1.0
Archived: 2026-10-07T20:57:12.486Z
SHA-256 of the archived text: 989a9f2bf45b5633227a413cd7fe757e5892f2d6d4e25886412600cab4fe209f
Stable URL of this page: https://metraipolis.com/documento/dc_2affcaec0e7c

How to cite

Vestrum: Improving Agent Harnesses by Adapting Their Verification, Structure and Memory — arXiv:2609.33822 — arXiv (en), https://arxiv.org/abs/2609.33822 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2affcaec0e7c (Metraipolis archive dc_2affcaec0e7c, sha256 989a9f2bf45b5633227a413cd7fe757e5892f2d6d4e25886412600cab4fe209f)

BibTeX

@misc{dc2affcaec0e7c,
  title = {Vestrum: Improving Agent Harnesses by Adapting Their Verification, Structure and Memory — arXiv:2609.33822},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33822},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2affcaec0e7c; sha256 989a9f2bf45b5633227a413cd7fe757e5892f2d6d4e25886412600cab4fe209f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33822. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

An agent harness controls how a language model accesses information, uses tools, preserves memory, and checks its work. Improving this software is costly when each evaluation requires a long interaction with an environment. We introduce Vestrum, a framework that turns failures in execution traces into scoped harness changes without training the task model. Its organizing overhypothesis is that tasks of a shared kind may exhibit recurring failures whose remedies transfer within that kind. Vestrum expresses failures as recognizable classes, proposes changes across verification, retrieval, decomposition, and knowledge synthesis, and screens their scope before evaluating them as a bundle. A persistent lessons file informs subsequent proposals. Across five settings and two baseline harnesses, the frozen harnesses improve held-out performance: UltraHorizon rises from 47.6 to 59.8 over GAM, Terminal-Bench 4 Hard from 63.7% to 70.3% of checks passed over Claude Code on eight held-out tasks at 1.03x test cost, and cell-type annotation agreement from 67.5% to 77.8% on held-out sections of one slide, alongside gains on LoCoMo and AMA-Bench. Across our searches, verification grounded in evidence helped both intermediate steps and final answers, at lower cost at intermediate steps, while critics asked to rebuild finished answers broke more than they repaired. On the three memory benchmarks, Vestrum also scores above the evaluated GEPA configurations in every paired evaluation.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.33822 — 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 2026-10-07T20:29:01.374Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).