Metraipolis › archive › dc_031a4c7db07d

Meta$^2$: Precomputing Reusable Knowledge for Online Harness Evolution in Long-Horizon Agents — arXiv:2611.05139

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.05139
License (as archived): CC0-1.0
Archived: 2026-11-04T18:35:26.960Z
SHA-256 of the archived text: 7517e2894487105e61045b07d399819b89a53d617f84c6199155069e5cc3d745
Stable URL of this page: https://metraipolis.com/documento/dc_031a4c7db07d

How to cite

Meta$^2$: Precomputing Reusable Knowledge for Online Harness Evolution in Long-Horizon Agents — arXiv:2611.05139 — arXiv (en), https://arxiv.org/abs/2611.05139 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_031a4c7db07d (Metraipolis archive dc_031a4c7db07d, sha256 7517e2894487105e61045b07d399819b89a53d617f84c6199155069e5cc3d745)

BibTeX

@misc{dc031a4c7db07d,
  title = {Meta$^2$: Precomputing Reusable Knowledge for Online Harness Evolution in Long-Horizon Agents — arXiv:2611.05139},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.05139},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_031a4c7db07d; sha256 7517e2894487105e61045b07d399819b89a53d617f84c6199155069e5cc3d745},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.05139. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Harness optimization improves long-horizon agents by refining the prompts, tools, memory, and hooks around a frozen model. Offline methods can learn from completed trajectories, but deploy one shared harness that may overfit training tasks and cannot react to evidence revealed during a new run. Online evolution can adapt before, during, or after one run through an evolver, but may have to reason about every intervention from scratch or rely on verifier feedback that is usually unavailable at test time. We introduce Meta$^2$, which moves offline optimization one level up: rather than optimizing a shared harness, it learns reusable knowledge for harness evolution and applies it within each task online. Offline, Meta$^2$ learns a textual Meta-Skill $M$ by contrasting training rollouts of the same task agent under three configurations: without an evolver, with an evolver guided by the current version of $M$, and with an evolver given privileged verifier information available only in training. Contrasts in trajectories and outcomes are aggregated across tasks to refine $M$ into reusable guidance on when and how to evolve, when to leave the harness unchanged, and which interventions to avoid. Online, an evolver combines the frozen $M$ with emerging trajectory evidence to decide whether to KEEP the harness unchanged or EVOLVE it through a task-local edit. Each edit applies to the rest of the same run without resetting its environment or progress. Thus, evolution guidance transfers across tasks while harness edits remain task-local, and held-out verifier feedback is never available to the evolver. On two long-horizon benchmarks, Terminal-Bench 2.1 and DeepSWE 1.1, Meta$^2$ achieves the highest performance among evaluated methods with GPT-5.6-Terra, reaching 81.00% and 44.44%, respectively. Ablations support the value of contrastive evolution guidance and keeping online harness edits task-local.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.05139 — 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 2026-11-04T18:32:37.355Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).