mlsanityData to 14 Oct 2026

Natural-Language Agent Harnesses

arXiv 2603.25723 ↗First posted 26 Mar 2026cs.CL cs.AIcited by 61

Not on the map. Why?

Cited by 41 in LLM agents and tool use, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Agent performance is strongly shaped by the surrounding harness: the external execution system around a model that organizes a task run. Yet this logic is usually buried in tightly coupled controller code, which makes harnesses hard to inspect, compare, transfer, and ablate. This paper asks whether the reusable design pattern of an agent harness can be represented as an executable natural-language object. We introduce Natural-Language Agent Harnesses (NLAHs), editable documents that describe run-level harness policy, and Intelligent Harness Runtime (IHR), a shared runtime that interprets these documents into agent calls, handoffs, state updates, validation gates, and artifact contracts. Across coding, terminal-use, and computer-use benchmarks, IHR-executed NLAHs achieve comparable task outcomes to code and prompted realizations, while exposing much shorter static harness policies. Module ablations further show that explicit harness modules are analyzable. These results suggest that agent harnesses can be turned from incidental glue around models into scientific representation objects.

Key facts. Cited by 61 papers in the map (data through 14 Oct 2026). Builds on 51 of its 85 references that are in the map. First posted on arXiv 26 Mar 2026.

Builds on

51 of 85 references.

References · 51

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  3. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,276
  4. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  5. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,373
  6. 2025s1: Simple test-time scalingcited by 1,341
  7. 2024OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environmentscited by 1,195
  8. 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,141
  9. 2026Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfacescited by 585
  10. 2024Executable Code Actions Elicit Better LLM Agentscited by 564
  11. 2024GPT-4V(ision) is a Generalist Web Agent, if Groundedcited by 563
  12. 2026SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learningcited by 278
38 more
  1. 2024Magentic-One: A Generalist Multi-Agent System for Solving Complex Taskscited by 232
  2. 2025AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agentscited by 223
  3. 2025Scaling Long-Horizon LLM Agent via Context-Foldingcited by 148
  4. 2022Prompting Is Programming: A Query Language for Large Language Modelscited by 110
  5. 2025Live-SWE-agent: Can Software Engineering Agents Self-Evolve on the Fly?cited by 95
  6. 2025Trae Agent: An LLM-based Agent for Software Engineering with Test-time Scalingcited by 94
  7. 2025MasRouter: Learning to Route LLMs for Multi-Agent Systemscited by 88
  8. 2024AutoFlow: Automated Workflow Generation for Large Language Model Agentscited by 83
  9. 2026AutoHarness: improving LLM agents by automatically synthesizing a code harnesscited by 76
  10. 2026Organizing, Orchestrating, and Benchmarking Agent Skills at Ecosystem Scalecited by 72
  11. 2026Meta Context Engineering via Agentic Skill Evolutioncited by 49
  12. 2026Skill-Pro: Learning Reusable Skills from Experience via Non-Parametric PPO for LLM Agentscited by 47
  13. 2024On the Worst Prompt Performance of Large Language Modelscited by 38
  14. 2025AMO-Bench: Large Language Models Still Struggle in High School Math Competitionscited by 38
  15. 2026When Single-Agent with Skills Replace Multi-Agent Systems and When They Failcited by 35
  16. 2026MAS-Orchestra: Understanding and Improving Multi-Agent Reasoning Through Holistic Orchestration and Controlled Benchmarkscited by 32
  17. 2026SkillCraft: Can LLM Agents Learn to Use Tools Skillfully?cited by 30
  18. 2025AnyMAC: Cascading Flexible Multi-Agent Collaboration via Next-Agent Predictioncited by 25
  19. 2025MathSmith: Towards Extremely Hard Mathematical Reasoning by Forging Synthetic Problems with a Reinforced Policycited by 24
  20. 2026ReCreate: Reasoning and Creating Domain Agents Driven by Experiencecited by 24
  21. 2024Perception Compressor: A Training-Free Prompt Compression Framework in Long Context Scenarioscited by 23
  22. 2025GMSA: Enhancing Context Compression via Group Merging and Layer Semantic Alignmentcited by 21
  23. 2026OS-Symphony: A Holistic Framework for Robust and Generalist Computer-Using Agentcited by 21
  24. 2024Prompts Are Programs Too! Understanding How Developers Build Software Containing Promptscited by 19
  25. 2024Do Advanced Language Models Eliminate the Need for Prompt Engineering in Software Engineering?cited by 19
  26. 2025FlowAgent: Achieving Compliance and Flexibility for Workflow Agentscited by 19
  27. 2026Building Effective AI Coding Agents for the Terminal: Scaffolding, Harness, Context Engineering, and Lessons Learnedcited by 18
  28. 2025MermaidFlow: Redefining Agentic Workflow Generation via Safety-Constrained Evolutionary Programmingcited by 16
  29. 2026OctoBench: Benchmarking Scaffold-Aware Instruction Following in Repository-Grounded Agentic Codingcited by 15
  30. 2025General Modular Harness for LLM Agents in Multi-Turn Gaming Environmentscited by 11
  31. 2026Read As Human: Compressing Context via Parallelizable Close Reading and Skimmingcited by 10
  32. 2024APPL: A Prompt Programming Language for Harmonious Integration of Programs and Large Language Model Promptscited by 8
  33. 2026U-Fold: Dynamic Intent-Aware Context Folding for User-Centric Agentscited by 8
  34. 2026AgentSpawn: Adaptive Multi-Agent Collaboration Through Dynamic Spawning for Long-Horizon Code Generationcited by 8
  35. 2025Promptware Engineering: Software Engineering for Prompt-Enabled Systemscited by 6
  36. 2026ContextCov: Deriving and Enforcing Executable Constraints from Agent Instruction Filescited by 6
  37. 2025Agint: Agentic Graph Compilation for Software Engineering Agentscited by 4
  38. 20263ViewSense: Spatial and Mental Perspective Reasoning from Orthographic Views in Vision-Language Modelscited by 4

50 most cited of 51 All 51 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 61.

61 papers.

Sort:

By area:

  1. 2026Code as Agent Harnesscited by 47
  2. 2026Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agentscited by 34
  3. 2026From Failed Trajectories to Reliable LLM Agents: Diagnosing and Repairing Harness Flawscited by 29
  4. 2026Who&When Pro: Can LLMs Really Attribute Failures in AI Agents?cited by 18
  5. 2026MemoHarness: Agent Harnesses That Learn from Experiencecited by 16
  6. 2026JIT-Agent: Scaling Harness Intelligence via Just-in-Time Harness Evolutioncited by 16
  7. 2026SafeHarness: Lifecycle-Integrated Security Architecture for LLM-based Agent Deploymentcited by 15
  8. 2026AutoSaddler: Automatic Harness Optimization with Durable Updates from Agent Execution Tracescited by 15
  9. 2026Agentic World Modeling: Foundations, Capabilities, Laws, and Beyondcited by 14
  10. 2026Harness Handbook: Making Evolving Agent Harnesses Readable,Navigable, and Editablecited by 13
  11. 2026From Question Answering to Task Completion: A Survey on Agent System and Harness Designcited by 12
  12. 2026Self-Evolving Embodied Agents via Skill-Harness Evolutioncited by 12
38 more
  1. 2026SEAGym: An Evaluation Environment for Self-Evolving LLM Agentscited by 11
  2. 2026Training with Harnesses: On-Policy Harness Self-Distillation for Complex Reasoningcited by 10
  3. 2026Crafter: A Multi-Agent Harness for Editable Scientific Figure Generation from Diverse Inputscited by 9
  4. 2026AutoPyVerifier: Learning Compact Executable Verifiers for Large Language Model Outputscited by 7
  5. 2026Remember the Decision, Not the Description: A Rate-Distortion Framework for Agent Memorycited by 7
  6. 2026SkillHarm: Lifecycle-Aware Skill-Based Attacks via Automated Constructioncited by 6
  7. 2026SkillGuard: A Permission-Centric Framework for Agent Skill Securitycited by 6
  8. 2026From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AIcited by 6
  9. 2026COMFYCLAW: Self-Evolving Skill Harnesses for Image Generation Workflowscited by 6
  10. 2026Dr-DCI: Scaling Direct Corpus Interaction via Dynamic Workspace Expansioncited by 5
  11. 2026Harness-Zero: Harness Distillation via Agent-as-Harnesscited by 5
  12. 2026HarnessBridge: Learnable Bidirectional Controller for LLM Agent Harnesscited by 3
  13. 2026From Agent Loops to Deterministic Graphs: Execution Lineage for Reproducible AI-Native Workcited by 2
  14. 2026Agent-ValueBench: A Comprehensive Benchmark for Evaluating Agent Valuescited by 2
  15. 2026Harness Engineering as Categorical Architecturecited by 2
  16. 2026Harnesses for Inference-Time Alignment over Execution Trajectoriescited by 2
  17. 2026AION: Next-Generation Tasks and Practical Harness for Time Seriescited by 2
  18. 2026Scaling Laws for Agent Harnesses via Effective Feedback Computecited by 2
  19. 2026Autonomous heterogeneous catalyst discovery with a self-evolving multi-agent digital twincited by 2
  20. 2026Orchestra-o1: Omnimodal Agent Orchestrationcited by 2
  21. 2026The Path to Self-Evolving Clinical Systems: Scaling Medical Agents from Assistance to Autonomycited by 2
  22. 2026SpatialGuard: Harness-Guided Verifiable Spatial Reasoning for Text-to-Image Generationcited by 2
  23. 2026EvoSafeHarness: Evolving Model- and Domain-Specific Harnesses for Securing Agentscited by 2
  24. 2026Agentic MIP Research: Accelerated Constraint Handler Generationcited by 1
  25. 2026OPD-Evolver: Cultivating Holistic Agent Evolver via On-Policy Distillationcited by 1
  26. 2026Harnessing Agent Skills: Architectural Patterns and a Reference Architecture for Skill-Mediated LLM Agentscited by 1
  27. 2026Harness the Memory: A Holistic Evaluation of Memory Substrates in Memory Agentscited by 1
  28. 2026Task-Aware Harness Provisioning for LLM Agents in Mission-Critical Infrastructure Operationscited by 1
  29. 2026CineForge: Self-Improving Agents for Long-Horizon Video Generationcited by 1
  30. 2026Agents in the Large: Perception-Centered Architecture for Persistent Agentscited by 1
  31. 2026WeAgent-MMGenEdit: A Full-Stack Recipe for Multimodal Agentic Image Generation and Editingcited by 1
  32. 2026When Malicious Instructions Persist: Persistent Memory Poisoning Attack on Harness-Based Agentscited by 1
  33. 2026Intermediate Artifacts as First-Class Citizens: A Data Model for Durable Intermediate Artifacts in Agentic Systemscited by 0
  34. 2026AI4BayesCode: From Natural Language Descriptions to Validated Modular Stateful Bayesian Samplerscited by 0
  35. 2026OCELOT: Inference-Leakage Budgets for Privacy-Preserving LLM Agentscited by 0
  36. 2026DS-Lighting: Making Agent Harnesses Explicit for Data-Science Automationcited by 0
  37. 2026Skill-based Agentic Evaluation for Real-time Data Science Taskscited by 0
  38. 2026The Work Behind Delegation: A Framework for Supervising AI Coding Agentscited by 0

50 most cited of 61 All 61 (JSON)