# Beyond Accuracy: Behavioral Dynamics of Agentic Multi-Hunk Repair

arXiv: https://arxiv.org/abs/2511.11012 · HTML: https://mlsanity.com/paper/2511.11012/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use.
- Builds on 25 of its 54 references that are in the map.
- First posted on arXiv 14 Nov 2025.

## Placed in

- LLM agents and tool use / [Software engineering coding agents](https://mlsanity.com/agents/software-engineering-coding-agents/)

## Builds on (most cited first, 25 of 25)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [OpenHands: An Open Platform for AI Software Developers as Generalist Agents](https://mlsanity.com/paper/2407.16741/) (2024)
- [RepairAgent: An Autonomous, LLM-Based Agent for Program Repair](https://mlsanity.com/paper/2403.17134/) (2024)
- [MAGIS: LLM-Based Multi-Agent Framework for GitHub Issue Resolution](https://mlsanity.com/paper/2403.17927/) (2024)
- [SWT-Bench: Testing and Validating Real-World Bug-Fixes with Code Agents](https://mlsanity.com/paper/2406.12952/) (2024)
- [Understanding Software Engineering Agents: A Study of Thought-Action-Result Trajectories](https://mlsanity.com/paper/2506.18824/) (2025)
- [SpecRover: Code Intent Extraction via LLMs](https://mlsanity.com/paper/2408.02232/) (2024)
- [MarsCode Agent: AI-native Automated Bug Fixing](https://mlsanity.com/paper/2409.00899/) (2024)
- [A Syntax-Guided Edit Decoder for Neural Program Repair](https://mlsanity.com/paper/2106.08253/) (2021)
- [Neural Program Repair with Execution-based Backpropagation](https://mlsanity.com/paper/2105.04123/) (2021)
- [You Name It, I Run It: An LLM Agent to Execute Tests of Arbitrary Projects](https://mlsanity.com/paper/2412.10133/) (2024)
- [Evaluating Agent-based Program Repair at Google](https://mlsanity.com/paper/2501.07531/) (2025)
- [Nopol: Automatic Repair of Conditional Statement Bugs in Java Programs](https://mlsanity.com/paper/1811.04211/) (2018)
- [Agentic Bug Reproduction for Effective Automated Program Repair at Google](https://mlsanity.com/paper/2502.01821/) (2025)
- [Beyond pip install: Evaluating LLM Agents for the Automated Installation of Python Projects](https://mlsanity.com/paper/2412.06294/) (2024)
- [Issue2Test: Generating Reproducing Test Cases from Issue Reports](https://mlsanity.com/paper/2503.16320/) (2025)
- [ITER: Iterative Neural Repair for Multi-Location Patches](https://mlsanity.com/paper/2304.12015/) (2023)
- [Harnessing Evolution for Multi-Hunk Program Repair](https://mlsanity.com/paper/1906.08903/) (2019)
- [RepairBench: Leaderboard of Frontier Models for Program Repair](https://mlsanity.com/paper/2409.18952/) (2024)
- [Characterizing Multi-Hunk Patches: Divergence, Proximity, and LLM Repair Challenges](https://mlsanity.com/paper/2506.04418/) (2025)
- [Contextual API Completion for Unseen Repositories Using LLMs](https://mlsanity.com/paper/2405.04600/) (2024)
- [A Controlled Experiment of Different Code Representations for Learning-Based Bug Repair](https://mlsanity.com/paper/2110.14081/) (2021)
- [Debug Smarter, Not Harder: AI Agents for Error Resolution in Computational Notebooks](https://mlsanity.com/paper/2410.14393/) (2024)
- [Katana: Dual Slicing-Based Context for Learning Bug Fixes](https://mlsanity.com/paper/2205.00180/) (2022)

## Built on by (most cited first, 2 of 2)

- [RepoRescue: An Empirical Study of LLM Agents on Whole-Repository Compatibility Rescue](https://mlsanity.com/paper/2607.01213/) (2026)
- [MultiFixer: A Coordinator-Proposer Based Multi-Agent Framework For Fixing Multi-Hunk Bugs](https://mlsanity.com/paper/2607.26591/) (2026)

## Abstract (arXiv)

Automated program repair has traditionally focused on single-hunk defects, overlooking multi-hunk bugs that are prevalent in real-world systems. Repairing these bugs requires coordinated edits across multiple, disjoint code regions, posing substantially greater challenges. We present the first systematic study of LLM-driven coding agents (Claude Code, Codex, Gemini-cli, and Qwen Code) on this task. We evaluate these four state-of-the-art agents on 404 multi-hunk bugs from the PolyHunk dataset, yielding 1,616 repair trajectories for large-scale behavioral analysis. We employ fine-grained metrics to assess localization, repair accuracy, regression behavior, and operational dynamics across agents. We find that localization capability varies substantially, with Codex achieving the highest success rate (75.3%) and Qwen Code the lowest (40.4%). Repair accuracy also differs widely, ranging from 26.98% (Qwen Code) to 92.82% (Claude Code), and consistently declines with increasing bug dispersion and complexity (hunk divergence and spatial proximity). High-performing agents (Claude Code and Codex) demonstrate superior semantic consistency, achieving positive average regression reduction, whereas lower-performing agents often introduce new test failures. Notably, agents do not fail fast; failed repairs consume substantially more resources (33%-440% more input tokens) and require longer execution time (35%-330%). Additionally, we developed Maple to provide agents with repository-level context. Empirical results show that Maple improves repair accuracy of Gemini-cli by ~21% through enhanced localization. By analyzing fine-grained metrics and trajectory-level analysis, this study moves beyond accuracy to explain how coding agents localize, reason, and act during multi-hunk repair. Our findings underscore the impact of bug divergence and spatial proximity on multi-hunk repair success for coding agents.