Agentic Harness for Real-World Compilers

2026-03-20

AI Summary

This work addresses the significant performance degradation—up to 60%—of large language models (LLMs) in compiler bug repair, a task hindered by cross-domain complexity and sparse bug reports. To bridge the gap between LLM capabilities and compiler engineering, the authors propose llvm-autofix, the first LLM-agent-oriented framework tailored for LLVM-based vulnerability repair. The framework integrates customized debugging and reproduction tools, introduces a reproducible benchmark suite named llvm-bench, and features a lightweight repair agent, llvm-autofix-mini. Experimental results demonstrate that the proposed approach improves repair success rates by approximately 22% over state-of-the-art methods, effectively enhancing LLMs’ applicability in the specialized domain of compiler infrastructure maintenance.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Compilers are critical to modern computing, yet fixing compiler bugs is difficult. While recent large language model (LLM) advancements enable automated bug repair, compiler bugs pose unique challenges due to their complexity, deep cross-domain expertise requirements, and sparse, non-descriptive bug reports, necessitating compiler-specific tools. To bridge the gap, we introduce llvm-autofix, the first agentic harness designed to assist LLM agents in understanding and fixing compiler bugs. Our focus is on LLVM, one of the most widely used compiler infrastructures. Central to llvm-autofix are agent-friendly LLVM tools, a benchmark llvm-bench of reproducible LLVM bugs, and a tailored minimal agent llvm-autofix-mini for fixing LLVM bugs. Our evaluation demonstrates a performance decline of 60% in frontier models when tackling compiler bugs compared with common software bugs. Our minimal agent llvm-autofix-mini also outperforms the state-of-the-art by approximately 22%. This emphasizes the necessity for specialized harnesses like ours to close the loop between LLMs and compiler engineering. We believe this work establishes a foundation for advancing LLM capabilities in complex systems like compilers. GitHub: https://github.com/dtcxzyw/llvm-autofix

Problem

• compiler bugs
• bug repair
• LLM
• complexity
• sparse bug reports

Innovation

• agentic harness
• compiler bug repair
• LLM for compilers
• llvm-autofix
• llvm-bench