The Harness Is the Model: Agent Scaffolding Changes a Fixed LLM's Compliance with Banking Conduct Rules, Unevenly Across Models
cs.CL, cs.AI, cs.CY
Submitted: 2026-11-04
Updated: 2026-11-04
Code: https://github.com/mhuyut/harness-is-the-model
Terminology
Sources
- From Confident Closing to Silent Failure: Characterizing False Success in LLM Agents
- From Prompts to Contracts: Harness Engineering for Auditable Enterprise LLM Agents
- Remembering More, Risking More: Longitudinal Safety Risks in Memory-Equipped LLM Agents
- General Agent Evaluation
- AgentDojo: A Dynamic Environment to Evaluate Prompt Injection Attacks and Defenses for LLM Agents
- Safety Under Scaffolding: How Evaluation Conditions Shape Measured Safety
- SABER: Benchmarking Operational Safety of LLM Coding Agents in Stateful Project Workspaces
- Large Language Models Cannot Self-Correct Reasoning Yet
- PolicyGuard: A Dialogue-Grounded Sub-Agent Verifier for Policy Adherence in LLM Agents
- AI Agents That Matter
- ST-WebAgentBench: A Benchmark for Evaluating Safety and Trustworthiness in Web Agents
- Auditing Agent Harness Safety
- Lost in the Middle: How Language Models Use Long Contexts
- ReguSim: Evaluating LLM Agent Rule Grounding in Financial Compliance
- LLM Evaluators Recognize and Favor Their Own Generations
- Large Language Models can Strategically Deceive their Users when Put Under Pressure
- Towards Understanding Sycophancy in Language Models
- The Cold-Start Safety Gap in LLM Agents
- Reviewer Capability Governs Rejection Targeting, Not Repair Skill: Evidence from LLM Execute-Review-Revise Pipelines
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data