Papers
arxiv:2609.39957

Learning When and How to Intervene: A Hindsight-Distilled Sentinel for Coding Agents

Published on Sep 30
Authors:
,
,
,

Abstract

A hindsight-distillation framework trains lightweight sentinels to decide pre-execution interventions and provide feedback, improving coding-agent task completion by preventing error propagation.

Coding agents solve repository-level tasks through sequences of actions, where a single erroneous action can misdirect subsequent decisions and increase recovery costs. Existing approaches use execution feedback for recovery or specialized checks to block errors, but deciding before execution whether intervention will benefit eventual task completion remains challenging. To address this challenge, we propose HiSentinel, a hindsight-distillation framework that trains lightweight 0.6B and 1.7B sentinels to select pre-execution interventions aimed at improving task completion rather than correcting every imperfect action. A privileged teacher uses recorded execution outcomes as evidence for intervention judgments, which are distilled into a causal student that receives only the pre-action context and proposed action. Beyond identifying whether and when to intervene, the sentinel must also provide actionable feedback that helps the coding agent recover or obtain necessary human input. To support these capabilities, we introduce SWE-Intervene, an action-level dataset constructed from software-engineering trajectories that annotates whether an action should be allowed, autonomously redirected, or paused for human assistance, together with corresponding intervention feedback. Across SWE-bench Verified Mini and Ask or Assume, HiSentinel consistently improves task completion across Sentinel scales and coding-agent families, with gains of up to 14% and 10%, respectively, while maintaining competitive token consumption. These results demonstrate that lightweight pre-execution intervention can effectively prevent error propagation and improve the reliability of autonomous coding agents.

Community

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2609.39957
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2609.39957 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2609.39957 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2609.39957 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.