Grow the Harness, Not the Context

A September 22, 2026 arXiv paper by Laizhen Li introduces Growing Harness, a failure-guided training paradigm that learns an LLM agent's harness as reusable executable code from a strategy-free scaffold, cutting LLM calls by 76.0-91.8% and deployed-agent inference cost by 74.4-98.6% versus a Tool-Calling agent. Across BrowseComp-Plus and WebArena-Verified with three deployment models from 4B to 120B parameters, Growing Harness achieved the highest mean success in five of six benchmark-model settings and trailed the best mean by 0.7 percentage points in the sixth. On WebArena-Verified its success held at 44.7-45.3% across model scales, while Tool-Calling fell to 6.7% with the 4B model.

Computer Science Artificial Intelligence Submitted on 22 Sep 2026 v1 https://arxiv.org/abs/2609.26760v1 , last revised 24 Sep 2026 this version, v2 Title:Grow the Harness, Not the Context: From Strategy-Free Scaffolds to Reusable Specialist Agents View PDF https://arxiv.org/pdf/2609.26760 HTML experimental https://arxiv.org/html/2609.26760v2 Abstract:Large language model LLM agents often handle streams of related tasks, yet standard harnesses repeatedly ask the model to reconstruct the same control decisions inside each task's context. We study whether task feedback can instead turn recurring control into reusable executable code, while reserving LLM calls for task-specific semantic reasoning. We introduce Growing Harness, a failure-guided training paradigm that learns the agent harness itself from a strategy-free scaffold that exposes fixed model and tool interfaces but encodes no task-solving controller. Function-level execution traces localize each failure to a bounded code surface, an optimizer repairs a window of failures jointly, and a success-first held-out gate rolls back repair sequences that harm prior capability. Accepted edits accumulate in one shared harness, allowing its control structure to emerge from task feedback. Across BrowseComp-Plus and WebArena-Verified with three deployment models from 4B to 120B parameters, Growing Harness achieves the highest mean success in five of six benchmark-model settings and trails the best mean by 0.7 pp. in the sixth. Relative to a Tool-Calling agent, it reduces LLM calls by 76.0-91.8% and deployed-agent inference cost by 74.4-98.6%. On WebArena-Verified, its success remains 44.7-45.3% across model scales, whereas Tool-Calling falls to 6.7% with the 4B model. Ablations show that trace-local edits, joint repair, and gate-based rollback each improve final success. These results show that persistent program growth can move recurring control out of model context and into low-cost code, yielding reusable specialist agents that remain effective with smaller deployment models. Submission history From: Laizhen Li view email https://arxiv.org/show-email/a718786d/2609.26760 Tue, 22 Sep 2026 17:40:45 UTC 353 KB \ v1\ https://arxiv.org/abs/2609.26760v1 v2 Thu, 24 Sep 2026 08:15:47 UTC 353 KB References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .