编码智能体演进现有代码时遵循先例还是指令？arXiv 新研究给出答案
Precedent or Intent? What Coding Agents Follow When Evolving Existing Code
AI 导读
arXiv 一项研究通过 5,625 次试验考察四个编码智能体在面对代码先例与新任务指令冲突时的选择。结果显示智能体压倒性地遵循仓库先例（359/360），但先例影响力取决于位置：三个本地支持文件带来 96% 采纳率，同级资源文件仅 4%。当任务意图与现有代码冲突时，书面优先级规则远胜代码示例（59/60 对 38/60），在开源仓库中加载规则可将联合成功率从 0/120 提升至 117/120。
正文 · 原文
Abstract:Existing code reveals past implementation decisions but rarely explains the reasoning behind them. When coding agents implement new features or modify existing functionality, past decisions may conflict with new task instructions. This raises a fundamental question for agentic software evolution: when multiple implementations satisfy the same functional requirements, how do coding agents decide whether to follow existing code or new task instructions? We investigate this question through bidirectional, controlled experiments in purpose-built repositories, followed by validation in real-world open-source repositories. Our approach isolates three competing sources of guidance: (1) agents' default preferences, (2) repository precedents, and (3) explicit task instructions. We conduct 5,625 trials across four coding agents and six implementation pairs, with validation in four open-source repositories spanning three programming languages. We find that agents share the same default implementation choices, yet overwhelmingly follow repository precedent when it conflicts with those defaults (359/360 trials). Crucially, not all precedents carry equal influence: three supporting local files yield 96% adoption, compared with only 4% for sibling-resource files, and changing file placement alone can shift adoption from 0% to 100% at a fixed example count. When new task intent conflicts with existing code, written priority rules achieve substantially higher success than code examples alone (59/60 versus 38/60 joint successes for two agents on two implementation pairs). The effectiveness of explicit rules also extends to open-source repositories, where a harness-loaded rule raises joint success from 0/120 to 117/120. These findings motivate design-decision traceability that preserves not only implementation choices but also the rationale and priorities that should guide future changes.
| Comments: | 16 pages, 7 figures, 6 tables |
| Subjects: | Software Engineering (cs.SE); Artificial Intelligence (cs.AI) |
| ACM classes: | D.2.7; I.2.2 |
| Cite as: | arXiv:2610.38244 [cs.SE] |
| (or arXiv:2610.38244v1 [cs.SE] for this version) | |
| https://doi.org/10.48550/arXiv.2610.38244 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Geyang Xu [view email]
[v1]
Wed, 28 Oct 2026 03:50:17 UTC (53 KB)
来源：arXiv cs.AI · arxiv.org