Summary This is the initial official release of IssueTrojanBench, accompanying the paper: "IssueTrojanBench: Benchmarking AI Coding Agents Against Malicious Issue Requests". What's Included: Core Benchmark Suite: All scripts for generating adversarial delivery vectors and perturbations. Evaluation Framework: Automated testing pipeline for coding agents (Claude Code, OpenHands, etc.). Dataset: The curated "seed" issues and adversarial variants used in our 696 experimental runs. Zenodo Archive This release is archived on Zenodo to provide a persistent DOI for academic citation.