Metraipolis › archive › dc_182ae320e078
Ecdysis: Efficient and Effective Training of Runtime Harnesses for LLM Agents — arXiv:2609.11677
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.11677
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T05:27:13.557Z
- SHA-256 of the archived text
c552c60b6cbcb527718cb96f078f2eb7c6c33fceb69464217273e3469a9730fa- Stable URL of this page
https://metraipolis.com/documento/dc_182ae320e078
How to cite
Ecdysis: Efficient and Effective Training of Runtime Harnesses for LLM Agents — arXiv:2609.11677 — arXiv (en), https://arxiv.org/abs/2609.11677 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_182ae320e078 (Metraipolis archive dc_182ae320e078, sha256 c552c60b6cbcb527718cb96f078f2eb7c6c33fceb69464217273e3469a9730fa)
BibTeX
@misc{dc182ae320e078,
title = {Ecdysis: Efficient and Effective Training of Runtime Harnesses for LLM Agents — arXiv:2609.11677},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.11677},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_182ae320e078; sha256 c552c60b6cbcb527718cb96f078f2eb7c6c33fceb69464217273e3469a9730fa},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.11677. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Self-evolving runtime harnesses can substantially improve the capabilities of large language model (LLM) agents and provide a promising paradigm for optimizing agent execution. Existing failure-driven approaches often treat observed agent failures as direct evidence for harness modification. A key challenge in failure-driven harness evolution is that observed failures can reflect either limitations of the underlying model or systematic deficiencies of the harness. Directly optimizing against individual failures can therefore induce model-specific accommodation and impair generalization across tasks and models. We study whether failure evidence accumulated across task instances can provide a more reliable signal for harness training. Our key insight is that failures recurring across distinct tasks provide stronger inductive evidence for systematic harness deficiencies than isolated failures. Based on this insight, we propose Ecdysis, which aggregates failure evidence across task instances before promoting recurring failure patterns into persistent harness evolution, biasing evolution toward repairs that are more likely to generalize beyond individual model behaviors. Ecdysis further employs collaborative failure analysis to refine modification specifications, trading additional evolution-time reasoning for improved modification quality. Across multiple LLMs and benchmarks, Ecdysis improves the reasoning accuracy of evolved harnesses by 18.56% over existing harness evolution while achieving up to 1.84x faster harness training. Ecdysis also enables more data-efficient training. Fine-grained analysis shows that Ecdysis reduces model-specific accommodation during evolution, while the resulting harnesses exhibit stronger cross-LLM generalization and lower inference-time token consumption. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-11 in cs.AI). Categorie arXiv: Software Engineering (cs.SE) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.11677 — 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 2026-10-16T05:17:52.185Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).