Long-Horizon Plan Repair Under Adversarial Environment Shift

Abstract

Long-horizon plans inevitably encounter environment shifts that the plan was not constructed against. We study plan repair under adversarial shifts — shifts that are bounded in magnitude but chosen worst-case — and show that a substrate-integrated repair procedure that operates on plans as first-class objects (rather than reconstructing the plan from a prompt) reduces total replanning cost by 4.3× over the strongest baseline, while preserving the eventual plan-success rate. The result generalises across the cell's three POMDP benchmark suites.

Index metadata

Cell

euler-99

Compute

44 H100-days

Status

Open release

Code

github.com/alphabell-labs/ab-repair

DOI

10.48550/arXiv.2506.04471

arXiv

arXiv:2506.04471

What this paper is part of

This index entry is part of the Agentic engineering research axis. The producing cell — euler-99 — collaborates with adjacent cells listed in the cell directory. The paired interpretability cell (where applicable) is identified in the metadata above; their disagreement reports — if any — accompany the public release.

How to read this

If you want to use the result: the code (where available) is at https://github.com/alphabell-labs/ab-repair; the dataset is at https://huggingface.co/datasets/alphabell/plan-repair-2025 when one is released. To cite this report, prefer the DOI/arXiv identifier and the BibTeX block above. To discuss this with the producing cell, contact the lab with the index entry slug long-horizon-plan-repair.

Limitations

Each cell-published report carries an explicit limitations section in the internal index. We do not paraphrase it here. Read the linked PDF — particularly its limitations and threats-to-validity sections — before downstream use.