Skip to content

Research1 min read

Optimizing LLM Tool-Agent Harnesses Without Retraining

Research demonstrates improving LLM tool agents by modifying the runtime harness – prompts and middleware – without retraining. This protocol measures and reports gains, highlighting the importance of reliable harness selection for multi-turn agents.

By OpenSmartRoute editorial · written through the router by writer-small

From arXiv cs.AI - “Beyond Prompts: Measuring and Optimizing LLM Tool-Agent Harnesses

A study investigates enhancing LLM tool agents by adjusting the runtime harness surrounding a fixed model. This approach focuses on prompt and tool-boundary middleware modifications, avoiding retraining. The protocol measures performance using mean and worst-condition held-out lift, repeatability, and RelLift95(B). It instantiates this protocol with prompt-only and prompt-plus-middleware optimizers, including PRISM, which identifies failures and directs repairs to prompts, middleware, or joint edit surfaces. On BFCL, tau2-Retail, and tau2-Telecom benchmarks, PRISM achieved mean held-out lifts of 14.2, 14.9, and 10.1 percentage points, respectively. The study attributes the primary margin to failure-surface routing and the edit-pattern constraint. Results indicate that some search procedures find large gains but may select brittle updates, emphasizing the need to report harness reliability alongside average held-out lift.

Source: https://arxiv.org/abs/2609.05736

Published Sep 9, 2026 · updated Sep 9, 2026 · 123 words

Keep reading

Related posts

More in Research

Research1 min read

ARC-Bench: Evaluating Action Ranking in Frozen JEPA World Models

ARC-Bench identifies a fundamental flaw in frozen JEPA-style world models: predicted action rankings based on latent space distances consistently fail to align with true cost. This research introduces a protocol to measure this failure and explains why it remains hidden due to replanning masking.