Agentic harness engineering improves agent capabilities with user-driven evolution
Agentic Harness Engineering Beta: Requirement-Centered Evolution of Agent Harnesses
Software EngineeringArtificial Intelligence
Summary
Improving AI agents often involves making changes based on test results, but this can miss important user needs and create patches that only work in specific cases. The authors created a new method called Agentic Harness Engineering Beta (AHE-Beta) that focuses on evolving AI agent systems based on real user requirements collected during use. This method uses a cycle of user feedback, team collaboration to fix issues broadly, and careful review to approve changes. Tests show AHE-Beta improves AI performance on tasks, uses less computing effort, and is preferred by human users compared to other methods.
What this means in practice
- •For ai development teams: Improve AI agent systems by evolving them based on real user needs and feedback rather than just benchmark scores, leading to broader and more effective improvements.
- •For software quality assurance teams: Integrate requirement-driven review and user feedback loops into automated system testing to catch implicit and interaction-based issues earlier and improve overall agent reliability.
Authors
Jiahang Lin, Jiahao Chen, Yiding Wang, Shichun Liu, Chengjun Pan, Shihan Dou, Honglin Guo, Yuyang Hu, Changhao Jiang, Zhenhua Han, Hang Yan, Qi Zhang, Tao Gui, Xuanjing Huang
Abstract
Real-world harness engineering is driven by explicit and implicit user requirements, which guide development and validation. However, existing harness evolution methods often overlook implicit requirements and interaction needs, while using a single evolution agent to modify the harness and benchmark scores to select changes. This process often produces narrow patches and leads to benchmark-specific local optima. We introduce Agentic Harness Engineering Beta (AHE-Beta), a requirement-centered loop that brings real-world engineering practices into harness evolution. The loop consists of three steps: (1) User-driven evaluation lets simulated users raise their requirements as issues while they work with the harness over multiple turns. (2) Team-based modification implements those requirements as generalizable changes rather than task-specific patches. (3) Requirement-driven selection merges a change only when it resolves a verified requirement and passes review. Empirically, on held-out tasks, AHE-Beta raises pass@1 from 21.9% to 33.3% while using 40.8% fewer tokens than Base Harness, and achieves higher pass@1 than Claude Code, Meta-Harness, and AHE. In a blind human A/B test, real users prefer AHE-Beta in 31 of 48 valid pairs and rate it higher on all five dimensions, matching the directional preference of the user simulator. The human A/B results show that AHE-Beta achieves its goal of aligning harness evolution with real-world user requirements. Without further evolution, the gains transfer to GLM-5.3 and DeepSeek-V4.1-Flash with pass@1 improvements of 15.8% and 5.3% over Base Harness. Together, these results show that requirement-centered harness evolution is a more general method for improving agent capabilities.