HarnessBridge: Learnable Bidirectional Controller for LLM Agent Harness

Hugging Face Daily Papers Papers

Summary

Introduces HarnessBridge, a learnable bidirectional controller that parameterizes the agent-environment interface for LLM agents, achieving performance comparable to specialized harnesses with reduced computational overhead on Terminal-Bench and SWE-bench.

Large language models are increasingly deployed as agents for long-horizon tasks, yet their performance is shaped not only by model capability and environment design, but also by the harness that mediates agent--environment interaction. Existing harnesses are largely manually engineered, making them difficult to scale as trajectories grow longer and interactions become more complex. In this work, we ask whether harness can be generated by a learnable plug-in module that can be trained in an end-to-end fashion. We introduce HarnessBridge, a lightweight learnable harness controller that parameterizes the agent--environment interface as a bidirectional projection. HarnessBridge learns two bidirectional projections: observation projection, which distills raw trajectories into compact, decision-relevant states, and action projection, which converts proposed actions into executable transitions or trajectory-grounded rejections. We train HarnessBridge on a harness supervision dataset via unified instruction tuning. On Terminal-Bench~2.0 and SWE-bench Verified, HarnessBridge matches or surpasses strong specialized harnesses while substantially reducing token usage and trajectory length, and generalizes from smaller generators to larger commercial models.
Original Article
View Cached Full Text

Cached at: 06/12/26, 02:52 AM

Paper page - HarnessBridge: Learnable Bidirectional Controller for LLM Agent Harness

Source: https://huggingface.co/papers/2606.12882

Abstract

Learnable harness controller called HarnessBridge is introduced to parameterize agent-environment interfaces through bidirectional projections, achieving performance comparable to specialized harnesses with reduced computational overhead.

Large language modelsare increasingly deployed as agents for long-horizon tasks, yet their performance is shaped not only by model capability and environment design, but also by the harness that mediates agent--environment interaction. Existing harnesses are largely manually engineered, making them difficult to scale as trajectories grow longer and interactions become more complex. In this work, we ask whether harness can be generated by a learnable plug-in module that can be trained in an end-to-end fashion. We introduce HarnessBridge, a lightweightlearnable harness controllerthat parameterizes the agent--environment interface as abidirectional projection. HarnessBridge learns twobidirectional projections:observation projection, which distills raw trajectories into compact, decision-relevant states, andaction projection, which converts proposed actions into executable transitions or trajectory-grounded rejections. We train HarnessBridge on a harness supervision dataset via unifiedinstruction tuning. OnTerminal-Bench~2.0 andSWE-benchVerified, HarnessBridge matches or surpasses strong specialized harnesses while substantially reducing token usage and trajectory length, and generalizes from smaller generators to larger commercial models.

View arXiv pageView PDFGitHub1Add to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.12882 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.12882 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.12882 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Harness Handbook: Making Evolving Agent Harnesses Readable,Navigable, and Editable

arXiv cs.AI

The Harness Handbook is a behavior-centric representation synthesized from agent harness codebases using static program analysis and LLM assistance, helping developers and coding agents locate code implementing specific behaviors. It introduces Behavior-Guided Progressive Disclosure (BGPD) to guide agents from high-level descriptions to relevant implementation details, improving localization accuracy and edit-plan quality.

Self-Harness: Harnesses That Improve Themselves

Hacker News Top

Self-Harness introduces a new paradigm where LLM-based agents iteratively improve their own operating harness by mining model-specific weaknesses, proposing harness modifications, and validating them through regression testing, achieving substantial performance gains on Terminal-Bench-2.0 across multiple base models.