What Does a Harness Buy? Tokens, Mostly
cs.AI, cs.LG, cs.SE
Submitted: 2026-10-03
Updated: 2026-10-03
Comments: 22 pages, 7 figures. Code and data: https://github.com/YangzeLiu/what-does-a-harness-buy
Code: https://github.com/YangzeLiu/what-does-a-harness-buy
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- The Scaffolding Matters More Than the Interface: A Controlled Comparison of MCP and CLI Tool Use Across Seven Agent Scaffoldings, Five Language Models, and One Software Task
- Don't Blame the Large Language Model: How Agent Harness Evolution Shapes Coding Agent Quality
- On Randomness in Agentic Evals
- An Empirical Study of Harness Design for Coding Agents
- Same Model, Different Harness: Different Coding-Agent Results
- Adding Error Bars to Evals: A Statistical Approach to Language Model Evaluations
- Inside the Scaffold: A Source-Code Taxonomy of Coding Agent Architectures
- Deployment Decision Reliability: A Generalizability-Theory Framework for Sizing Long-Horizon Agent Evaluations
- Scaffold Effects on GAIA: A Controlled Comparison
- The Scaffold Effect in Coding Agents: Harness Choice as a Hidden Variable in Coding-Agent Evaluation
- When Does Restricting a Coding Agent to execute_code Help? A Regime $\times$ Agent-Design Ablation
- Harness-Bench: Evaluating Agent Harnesses across Models and Realistic Workflows
- Stop Comparing LLM Agents Without Disclosing the Harness
- Claw-SWE-Bench: A Benchmark for Evaluating OpenClaw-style Agent Harnesses on Coding Tasks
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems