Metraipolis › archive › dc_c52a697229a4

An Empirical Study of Harness Design for Coding Agents — arXiv:2609.20804

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.20804
License (as archived)
CC0-1.0
Archived
2026-10-10T03:22:23.849Z
SHA-256 of the archived text
24ab8598dc810c4cde2540b844d4cda75df7813970b7ec602d2403fb366e8c02
Stable URL of this page
https://metraipolis.com/documento/dc_c52a697229a4

How to cite

An Empirical Study of Harness Design for Coding Agents — arXiv:2609.20804 — arXiv (en), https://arxiv.org/abs/2609.20804 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c52a697229a4 (Metraipolis archive dc_c52a697229a4, sha256 24ab8598dc810c4cde2540b844d4cda75df7813970b7ec602d2403fb366e8c02)
BibTeX
@misc{dcc52a697229a4,
  title = {An Empirical Study of Harness Design for Coding Agents — arXiv:2609.20804},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.20804},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c52a697229a4; sha256 24ab8598dc810c4cde2540b844d4cda75df7813970b7ec602d2403fb366e8c02},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.20804. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Coding harnesses shape how autonomous coding agents translate model capabilities into long-horizon software-engineering performance, yet existing work typically evaluates harnesses as monolithic systems, leaving the effectiveness of individual components unclear. To enable component-level comparisons, we study this question with a lightweight coding harness whose execution loop is fixed while three components are varied: planning, action space, and context management. Across four models evaluated on SWE-Bench Verified and Terminal-Bench 2.1, we evaluate 176 matched settings spanning five context-management strategies, four context-window budgets, and targeted ablations of planning and action space. We find that: (1) Context management becomes increasingly valuable as the context-window budget tightens, with most of its benefit coming from preventing context-overflow failures. (2) Staging rule-based elision before LLM-based summarization provides the strongest overall efficiency among the context-management strategies, whereas making elided content recoverable adds machinery that models rarely use and yields no accuracy gain. (3) Planning shifts from an accuracy scaffold for weaker models to a cost saver for stronger models, with little change in accuracy. (4) Predefined tools improve performance for models with weaker bash proficiency, whereas bash-capable models can operate effectively with a bash-only interface and achieve substantially lower cost, especially on command-line-centric tasks. Trajectory-level analysis explains these effects: context management extends execution trajectories without substantially altering agent behavior, planning changes where trajectories stop, and the action space changes the granularity at which code is written. These findings inform model- and budget-aware harness design and provide a modular framework for evaluating future harness components.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-18 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Machine Learning (cs.LG); Software Engineering (cs.SE).
Pagina del lavoro: https://arxiv.org/abs/2609.20804 — 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 2026-10-10T03:04:13.763Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).