# PatchHolmes: Agentic Patch Retrieval via Listwise Selection

arXiv: https://arxiv.org/abs/2609.38807 · HTML: https://mlsanity.com/paper/2609.38807/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use.
- Builds on 17 of its 27 references that are in the map.
- First posted on arXiv 30 Sep 2026.

## Placed in

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

## Builds on (most cited first, 17 of 17)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection](https://mlsanity.com/paper/2310.11511/) (2023)
- [Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning](https://mlsanity.com/paper/2503.09516/) (2025)
- [Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models](https://mlsanity.com/paper/2506.05176/) (2025)
- [Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions](https://mlsanity.com/paper/2212.10509/) (2022)
- [BrowseComp-Plus: A More Fair and Transparent Evaluation Benchmark of Deep-Research Agent](https://mlsanity.com/paper/2508.06600/) (2025)
- [FlashRAG: A Modular Toolkit for Efficient Retrieval-Augmented Generation Research](https://mlsanity.com/paper/2405.13576/) (2024)
- [BRIGHT: A Realistic and Challenging Benchmark for Reasoning-Intensive Retrieval](https://mlsanity.com/paper/2407.12883/) (2024)
- [Code Researcher: Deep Research Agent for Large Systems Code and Commit History](https://mlsanity.com/paper/2506.11060/) (2025)
- [Tool-integrated Reinforcement Learning for Repo Deep Search](https://mlsanity.com/paper/2508.03012/) (2025)
- [RANGER -- Repository-Level Agent for Graph-Enhanced Retrieval](https://mlsanity.com/paper/2509.25257/) (2025)
- [The Landscape of Toxicity: An Empirical Investigation of Toxicity on GitHub](https://mlsanity.com/paper/2502.08238/) (2025)
- [Incivility in Open Source Projects: A Comprehensive Annotated Dataset of Locked GitHub Issue Threads](https://mlsanity.com/paper/2402.04183/) (2024)
- [Can Highlighting Help GitHub Maintainers Track Security Fixes?](https://mlsanity.com/paper/2411.11646/) (2024)
- [SPFinder: Improving the Context Length and Scalability for Tracing Known Vulnerability Patches](https://mlsanity.com/paper/2503.22935/) (2025)
- [Favia: Forensic Agent for Vulnerability-fix Identification and Analysis](https://mlsanity.com/paper/2602.12500/) (2026)

## Abstract (arXiv)

Patch retrieval, the task of finding the commit that fixes a known vulnerability, is the foundation of vulnerability management workflows, yet 60% to 63% of CVEs in the major advisory databases lack a patch link. We present PatchHolmes, a two-phase patch retrieval system that pairs a hybrid first-stage retriever with an agentic second-stage inspection loop. Unlike pointwise prior work that scores each candidate independently, the Phase 2 agent reads the top-100 listwise: it sees the full candidate list at once and selectively reads 3 to 10 commits through four budgeted tools before submitting a single best commit. On GitHubAD, PatchHolmes beats the pointwise binary classifier Favia by 25.34% Recall@1 and the retrieve-and-CoT baseline IRCoT by 31.40%, at one agent conversation per CVE versus Favia's ten; with the candidate set held identical, the agent adds 27.32% Recall@1 over taking the retriever's top candidate, and the same agent, transferred unchanged to PatchFinder_top10, lifts Recall@1 from PatchFinder's own top-1 pick (24.28%) to 39.86%. Swapping the LLM backbone within the Qwen family changes Recall@1 by under 1%, and a second model family (gpt-oss) stays far above the no-agent floor, so the gain comes from the listwise agent loop; the entire system runs on a frozen open-weight model over a local Git repository, without fine-tuning or external search APIs.