Artifact of the paper "Dependency-aware Residual Risk Analysis"

This repository contains the replication package and analysis artifacts for the paper “Dependency-aware Residual Risk Analysis”, accepted at the 48th International Conference on Software Engineering (ICSE 2026). It enables reproduction of all experimental results reported in the paper and provides the tooling required to run the experiments on new systems. Contents Overview Replication materials: Docker-based setup and scripts for running blackbox and greybox fuzzing experiments built on top of AFL++. Experimental data: Collected fuzzing results and intermediate data used for residual risk and discovery probability analysis. Analysis scripts: Jupyter notebooks for reproducing all figures and tables corresponding to the paper’s research questions (RQ1–RQ4). Paper PDF: A copy of the ICSE 2026 paper is included for reference. Please read README.md for the entire artifact structure and replication/Experiment-Reproduction.md to reproduce the experiment. Experiment Scope The experiments focus on blackbox and greybox fuzzing with residual risk estimation. It includes both the single-fuzzing-run Dockerfile and how to expand the experiment by combining it with fuzzbench, which generates the data for the paper's evaluation. Our AFL++-based instrumentation records: discovered basic blocks, singleton basic blocks (executed exactly once), singleton clusters (as defined in the paper), which are then used to estimate discovery probability and residual risk over time. Detailed instructions for reproducing the fuzzing experiments are provided in replication/Experiment-Reproduction.md. Analysis Reproduction For each research question, a dedicated Jupyter notebook reproduces the corresponding results: RQ1: Discovery probability estimation and stop-time accuracy RQ2: Node reduction effects and fuzzing performance RQ3: Residual risk estimation accuracy RQ4: Discovery probability estimation under greybox fuzzing Running the notebooks on the provided data regenerates the figures and tables reported in the paper. Citation If you use this artifact, please cite the ICSE 2026 paper. Citation information and BibTeX are provided in CITATION.cff.