AI ethics evidence index Search
Source-linked daily record
Evidence record 16234 · automatically gathered

Model or Harness? An Interaction-Centric Taxonomy for Localizing Agent Failures

Existing evaluations often reduce agent failures to system-level outcomes, obscuring where the fault originated and which intervention would improve the agent system. This creates a repair-assignment problem: the same visible failure may call for model post-training, harness engineering, environment redesign, or benchmark repair depending on its source. Because agent behavior emerges from interactions among models, harnesses, users, tools, memory, and environments, outcome-level labels are often

Record details

Published: 29 July 2026
Source: HuggingFace Daily Papers
Category: Research
Topics: Agents & autonomy · Environment
Retrieved: 5 August 2026

source-onlyevidence status

These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.

Shared topic: Agents & autonomy · Shared topic: Environment

Echoverse: Deep, Evolving Environments for Training Computer-Use Agents at Scale

HuggingFace Daily Papers · 29 July 2026

Shared topic: Agents & autonomy · Shared topic: Environment

Harness-G: A Graph-Structured Harness for Search Agents

HuggingFace Daily Papers · 29 July 2026

Shared topic: Agents & autonomy · Shared topic: Environment

BioVLN: A Simulation Platform for Visual Language Navigation in Biomedical Laboratories

arXiv cs.AI · 29 July 2026

Shared topic: Agents & autonomy · Shared topic: Environment

SecRespond: Benchmarking AI Agents for Real-World Post-Compromise Incident Response

arXiv cs.AI · 29 July 2026

Shared topic: Agents & autonomy · Shared topic: Environment

Rethinking Self-Evolution: A Constrained Exploration-Exploitation Process for Mitigating Skill Overfitting

arXiv · 29 July 2026

Shared topic: Agents & autonomy · Shared topic: Environment

Eco3S: Complex Socio-Economic System Simulation via Agent-Based Models

arXiv · 29 July 2026

How to cite this record

ethics.ai (29 July 2026), “Model or Harness? An Interaction-Centric Taxonomy for Localizing Agent Failures,” evidence record 16234, https://ethics.ai/record/16234 (originally published by HuggingFace Daily Papers).

JSON

Use and limitations

This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.