Metraipolis › archive › dc_952a5479ddff
One Readout, Many Repairs: Diffusion-Guided Hierarchical Search for Tool-Agent Repair — arXiv:2609.34879
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.34879
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T20:58:16.976Z
- SHA-256 of the archived text
e7b07e16db45c0581f2b7b93ff536681cc01bcd685bf9c19d05b08258f53cbf6- Stable URL of this page
https://metraipolis.com/documento/dc_952a5479ddff
How to cite
One Readout, Many Repairs: Diffusion-Guided Hierarchical Search for Tool-Agent Repair — arXiv:2609.34879 — arXiv (en), https://arxiv.org/abs/2609.34879 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_952a5479ddff (Metraipolis archive dc_952a5479ddff, sha256 e7b07e16db45c0581f2b7b93ff536681cc01bcd685bf9c19d05b08258f53cbf6)
BibTeX
@misc{dc952a5479ddff,
title = {One Readout, Many Repairs: Diffusion-Guided Hierarchical Search for Tool-Agent Repair — arXiv:2609.34879},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.34879},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_952a5479ddff; sha256 e7b07e16db45c0581f2b7b93ff536681cc01bcd685bf9c19d05b08258f53cbf6},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34879. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Tool agents use large language models to act through external tools, yet successfully executed calls can still leave user requests unfulfilled. Tool-agent repair seeks alternative call sequences that execute successfully and fulfill the original requests. However, repair requires exploring both operation choices and their concrete realizations, making complete-sequence regeneration costly. Moreover, regeneration repeats operation selection even when failure arises from how those operations are realized. The resulting challenge is to reduce this repetition while preserving exploration of alternative operations and realizations. Therefore, we formulate repair as hierarchical search over operation supports, which we introduce as sets of permitted operation types that define reusable search regions for concrete tool-call sequences. We propose ReCommit, a training-free, diffusion-guided framework for improving tool-agent failure recovery while reducing repair computation. ReCommit amortizes operation-level proposal computation across repair trials by reusing operation-type scores from a single parallel readout of a masked diffusion language model. These scores guide search across supports, while realization search explores alternative entity bindings, arguments, and action composition within each support. Experiments on real failures across four enterprise services in the Agent-Diff benchmark show 75.9\% and 63.2\% relative recovery gains with 61.3\% and 51.3\% reductions in mean full-budget repair time at repair budgets $B=3$ and $B=13$, respectively, over the strongest evaluated 8B comparison method. ReCommit achieves a favorable recovery--cost trade-off, including in comparisons with the evaluated 32B models. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.34879 — 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 2026-10-07T20:29:01.700Z 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).