Search papers, labs, and topics across Lattice.
This paper introduces AutoSaddler, an automatic harness optimization framework that enhances the reliability of LLM agents on long-horizon tasks by leveraging failure signals from agent execution traces. By treating harness improvement as an offline learning problem, AutoSaddler iteratively updates harnesses through structured patch generation and validation-based selection, leading to significant performance gains across multiple benchmarks. The results indicate that deep debugging, targeted modifications, and generalization-aware selection are critical for effective harness optimization, highlighting a promising avenue for improving agent robustness.
AutoSaddler achieves up to 10% performance improvement in LLM agents by automatically optimizing harnesses based on execution failure signals.
LLM agents remain unreliable on long-horizon tasks, where small local failures can compound over extended interactions and lead to overall task failure. Although external harnesses can substantially improve robustness, harness design remains a manual and expensive process that requires searching over a large space of prompts, tool configurations, and control logic. We propose AutoSaddler, an automatic harness optimization framework that formulates harness improvement as an offline learning problem and iteratively updates the harness using failure signals from mini-batches. AutoSaddler combines failure-trace diagnosis, structured patch generation that treats the harness as code, and validation-based update selection. Experiments on GAIA2, SWE-Bench Pro, and Terminal-Bench 2.0 show that AutoSaddler substantially improves agent performance over the corresponding base harnesses, achieving gains of 9.0, 9.6, and 10.0 percentage points, respectively. Ablation studies further suggest that effective harness optimization benefits from three ingredients: deep debugging rather than shallow reflection, targeted modifications rather than unconstrained editing, and generalization-aware selection rather than trajectory-specific repair. Together, these results suggest that automatic harness optimization is a promising path toward more performant and reliable agent systems.
Microsoft Research