Search papers, labs, and topics across Lattice.
This paper evaluates and improves clarification-seeking abilities of LLM agents in software engineering tasks with underspecified instructions. They introduce an uncertainty-aware multi-agent system that decouples underspecification detection from code execution, using OpenHands and Claude Sonnet 4.5. The proposed system significantly outperforms single-agent setups on a modified SWE-bench Verified benchmark, achieving a 69.40% task resolve rate and demonstrating well-calibrated uncertainty in query generation.
LLMs can become proactive collaborators that independently recognize when to ask questions to elicit missing information in real-world, underspecified tasks.
As Large Language Model (LLM) agents are increasingly deployed in open-ended domains like software engineering, they frequently encounter underspecified instructions that lack crucial context. While human developers naturally resolve underspecification by asking clarifying questions, current agents are largely optimized for autonomous execution. In this work, we systematically evaluate the clarification-seeking abilities of LLM agents on an underspecified variant of SWE-bench Verified. We propose an uncertainty-aware multi-agent scaffold that explicitly decouples underspecification detection from code execution. Our results demonstrate that this multi-agent system using OpenHands + Claude Sonnet 4.5 achieves a 69.40% task resolve rate, significantly outperforming a standard single-agent setup (61.20%) and closing the performance gap with agents operating on fully specified instructions. Furthermore, we find that the multi-agent system exhibits well-calibrated uncertainty, conserving queries on simple tasks while proactively seeking information on more complex issues. These findings indicate that current models can be turned into proactive collaborators, where agents independently recognize when to ask questions to elicit missing information in real-world, underspecified tasks.