AgentCompass: A Unified Evaluation Infrastructure for Agent Capabilities
Abstract
As Large Language Models (LLMs) evolve into autonomous agents, the need for unified evaluation infrastructure becomes critical. However, current evaluation pipelines remain highly fragmented and tightly coupled, hindering reproducibility and causing redundant engineering. To address this, we introduce AgentCompass, an open-source, lightweight, and extensible infrastructure for evaluating LLM-based agents. AgentCompass organizes the evaluation process around three independent components, namely Benchmark, Harness, and Environment, thereby enabling flexible configurations without requiring the reimplementation of complex execution logic. Furthermore, it features a fault-tolerant asynchronous runtime and comprehensive trajectory analysis tools to transparently diagnose nuanced failure modes like reward-hacking. Natively supporting over 20 benchmarks across five capability dimensions, AgentCompass provides the community with a scalable and reproducible infrastructure for advancing agent research.
Cite
@article{arxiv.2607.13705,
title = {AgentCompass: A Unified Evaluation Infrastructure for Agent Capabilities},
author = {Zichen Ding and Jiaye Ge and Shufan Jiang and Kai Chen and Mo Li and Qingqiu Li and Zehao Li and Zonglin Li and Tiaohao Liang and Shudong Liu and Zerun Ma and Zixing Shang and Wenhui Tian and Zun Wang and Liwei Wu and Zhenyu Wu and Jun Xu and Bowen Yang and Dingbo Yuan and Qi Zhang and Songyang Zhang and Peiheng Zhou and Dongsheng Zhu},
journal= {arXiv preprint arXiv:2607.13705},
year = {2026}
}