XRepair - Unifying Retrieval, Repair, and Evaluation for Explainable LLM-Based Vulnerability Fixes
Bibliographic record
Abstract
This paper presents XRepair, a web-based tool for automated and explainable software vulnerability repair. XRepair integrates retrieval, generation, and evaluation into a single interactive framework. It augments large language models (LLMs) with knowledge retrieved from CWE/CVE datasets through both embedding-based and graph-based strategies. XRepair enables repairs that are contextualized with vulnerability metadata and related examples. The tool provides end-to-end functionalities, including users inputting a CWE/CVE identifier and vulnerable code snippet, configuring the retrieval strategy and LLM, and receiving repair recommendations. Unlike patch-only systems, the recommendations consist of an issue description, explicit repair guidance, and concrete code diffs. To assess repair quality, XRepair incorporates an LLM-as-a-Judge evaluation framework that scores outputs across five criteria, namely relevance, completeness, correctness, identification of vulnerable code, and guidance. This design supports reproducible comparison of retrieval strategies and produces transparent recommendations. Built using both the strategy and factory design patterns, XRepair allows new retrieval algorithms and data sources to be added. The tool and documentation are publicly available with a tool for comparative evaluation and practical application of explainable vulnerability repairs. The demo for XRepair can be accessed at https://youtu.be/5IcGzNJXAY.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.015 | 0.069 |
| Meta-epidemiology (narrow) | 0.003 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.003 |
| Bibliometrics | 0.007 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.005 | 0.007 |
| Open science | 0.004 | 0.008 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.014 | 0.007 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".