Metraipolis › archive › dc_f99c72f90be3
ModularRSI: Modular and Generalizable Recursive Harness Self-Improvement — arXiv:2609.14857
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.14857
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T23:15:34.629Z
- SHA-256 of the archived text
ec1375d0776228366cb7afcbfcb3a2784a3dd509675d349dce4c54453797cf3a- Stable URL of this page
https://metraipolis.com/documento/dc_f99c72f90be3
How to cite
ModularRSI: Modular and Generalizable Recursive Harness Self-Improvement — arXiv:2609.14857 — arXiv (en), https://arxiv.org/abs/2609.14857 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f99c72f90be3 (Metraipolis archive dc_f99c72f90be3, sha256 ec1375d0776228366cb7afcbfcb3a2784a3dd509675d349dce4c54453797cf3a)
BibTeX
@misc{dcf99c72f90be3,
title = {ModularRSI: Modular and Generalizable Recursive Harness Self-Improvement — arXiv:2609.14857},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.14857},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f99c72f90be3; sha256 ec1375d0776228366cb7afcbfcb3a2784a3dd509675d349dce4c54453797cf3a},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.14857. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Recent work extends recursive self-improvement (RSI) to agent harnesses for long-horizon coding and terminal tasks, enabling agents to improve execution mechanisms from experience. However, generalizable harness RSI remains challenging. First, evolving harnesses on evaluation benchmarks or their subsets makes it difficult to distinguish reusable improvements from benchmark-specific adaptation. Second, single-trajectory updates can conflate systematic harness deficiencies with instance-specific reasoning and solution details, producing modifications that transfer poorly to unseen tasks. Third, localizing recurring behavioral deficiencies within monolithic harnesses is difficult, while whole-harness optimization can entangle unrelated mechanisms and complicate attribution and validation. We propose ModularRSI, a benchmark-disjoint, contrastive, and modular framework for generalizable harness evolution. ModularRSI contrasts successful and failed trajectories for the same task and aggregates evidence across tasks to identify recurring behavioral deficiencies. It decomposes the evolvable harness into five functional modules: Agent Loop, Tool Use, Observation Management, Context Management, and Task Completion Detection. Each module evolves independently within a restricted modification scope, followed by an integration stage that combines the evolved modules into a unified harness and resolves potential conflicts. To support benchmark-disjoint evolution, we curate 2,000 executable evolution tasks from external sources that are disjoint from downstream evaluation benchmarks. Experiments on TB2.0 and SWE-Bench Verified show consistent improvements on unseen in-domain and cross-domain tasks, with the evolved harness also transferring across different foundation models. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-15 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.14857 — 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 2026-10-15T22:51:09.911Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).