Counterfactual Trajectory Replay for Off-Policy Agent Debugging

Abstract

Debugging long-horizon agents is hard precisely because the failure that motivates the debug session may have arisen tens of thousands of steps before the visible symptom. We introduce trajectory-replay, a debugging mode built on the alphabell substrate's content-addressed traces, in which a deviation at any prior step can be simulated forward under the same model weights and substrate state. We report on six cell-internal debugging sessions where the technique uncovered root causes that the producing cell had been unable to find through prompt-level inspection, and discuss the operational requirements for replay to be both fast and faithful.

Index metadata

Cell

fourier-67 / babbage-14

Compute

22 H100-days (analysis only)

Status

Open release

Code

github.com/alphabell-labs/ab-replay

Substrate version

v1.4+

DOI

10.48550/arXiv.2512.00417

arXiv

arXiv:2512.00417

What this paper is part of

This index entry is part of the Agentic engineering research axis. The producing cell — fourier-67 — 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-replay; the dataset is at TBD 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 counterfactual-trajectory-replay.

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.