Externalizing Requirement-to-Repair Artifacts as Observable Traces for LLM-Based Program Repair
cs.AI, cs.SE
Submitted: 2026-09-14
Updated: 2026-09-14
Comments: Accepted at REALM Workshop, EMNLP 2026. Code: https://github.com/ForeverMJ/THEMIS
Code: https://github.com/ForeverMJ/THEMIS
License: http://creativecommons.org/licenses/by/4.0/
The gist: Repository-level repair requires not only correct patches but also inspectable records that explain how issue requirements are translated into code changes and post-edit evidence.
Terminology
Abstract
Repository-level repair requires not only correct patches but also inspectable records that explain how issue requirements are translated into code changes and post-edit evidence. We contribute THEMIS, a stage-aware repair workflow that externalizes this requirement-to-repair process through semantic interpretation, a runtime requirement-code graph, graph-derived Developer guidance, retained repair rationale and patches, and post-edit audit records. A retrospective audit of 300 SWE-bench Lite cases demonstrates that these artifacts provide broad support for cross-stage inspection: a complete Developer rationale is available for 288 cases, and 214 cases (71.3%) retain a complete audited field set connecting the selected stages. The retained records further enable systematic measurement of cross-stage correspondence: target symbols recur in 62.6% of Developer rationales and in 62.8% of patches, rising to 75.8% when related symbols are included. In a paired 100-case comparison, the relational workflow resolves 19 cases versus 9 for the direct same-input condition; because the two conditions also differ in Analyzer output, graph-derived distillation, and Judge records, we report this as preliminary, workflow-level evidence rather than a causal effect of the graph component. Together, these results show that THEMIS makes otherwise implicit requirement-to-repair transitions inspectable, enabling systematic examination of how repair decisions persist, align, and evolve across stages.
Sources
- Learning to Represent Programs with Graphs
- code2seq: Generating Sequences from Structured Representations of Code
- code2vec: Learning Distributed Representations of Code
- GraphCodeBERT: Pre-training Code Representations with Data Flow
- SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
- An Analysis of the Automatic Bug Fixing Performance of ChatGPT
- From Agent Traces to Trust: A Survey of Evidence Tracing and Execution Provenance in LLM Agents
- Agentless: Demystifying LLM-based Software Engineering Agents
- Conversational Automated Program Repair
- SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems
- VSAL: A Vision Solver with Adaptive Layouts for Graph Property Detection