RubricRefine: Improving Tool-Use Agent Reliability with Training-Free Pre-Execution Refinement

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: LeVine, Will, Evers, Brendan, Saltwick, Sam, Venkatesh, Abhay
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913133273022464
author LeVine, Will
Evers, Brendan
Saltwick, Sam
Venkatesh, Abhay
author_facet LeVine, Will
Evers, Brendan
Saltwick, Sam
Venkatesh, Abhay
contents Iterative self-refinement is a popular inference-time reliability technique, but its effectiveness in code-mode tool use depends heavily on the structure of the feedback signal: unstructured critique helps inconsistently across models, and even revision with real execution feedback improves only modestly ($0.75$ vs. $0.65$ baseline). The dominant failures are inter-tool contract violations (wrong output shape, incorrect tool routing, broken argument provenance) that run to completion without raising errors, making runtime feedback insufficient. We introduce RubricRefine, a training-free method for pre-execution semantic contract verification that generates task- and registry-specific rubrics, scores candidate code against explicit contract checks, and iteratively repairs failures before any execution occurs. RubricRefine reaches $0.86$, averaged across seven models, on M3ToolEval with zero execution attempts, improving over prior inference-time baselines with up to $2.6\times$ lower latency. Performance remains flat on the predominantly single-step API-Bank, consistent with the method's reliance on inter-tool contract structure. A rubric-category ablation and calibration analysis further characterize when and why the method works.
format Preprint
id arxiv_https___arxiv_org_abs_2605_09730
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle RubricRefine: Improving Tool-Use Agent Reliability with Training-Free Pre-Execution Refinement
LeVine, Will
Evers, Brendan
Saltwick, Sam
Venkatesh, Abhay
Machine Learning
Software Engineering
Iterative self-refinement is a popular inference-time reliability technique, but its effectiveness in code-mode tool use depends heavily on the structure of the feedback signal: unstructured critique helps inconsistently across models, and even revision with real execution feedback improves only modestly ($0.75$ vs. $0.65$ baseline). The dominant failures are inter-tool contract violations (wrong output shape, incorrect tool routing, broken argument provenance) that run to completion without raising errors, making runtime feedback insufficient. We introduce RubricRefine, a training-free method for pre-execution semantic contract verification that generates task- and registry-specific rubrics, scores candidate code against explicit contract checks, and iteratively repairs failures before any execution occurs. RubricRefine reaches $0.86$, averaged across seven models, on M3ToolEval with zero execution attempts, improving over prior inference-time baselines with up to $2.6\times$ lower latency. Performance remains flat on the predominantly single-step API-Bank, consistent with the method's reliance on inter-tool contract structure. A rubric-category ablation and calibration analysis further characterize when and why the method works.
title RubricRefine: Improving Tool-Use Agent Reliability with Training-Free Pre-Execution Refinement
topic Machine Learning
Software Engineering
url https://arxiv.org/abs/2605.09730