Effective harnesses for long-running agents

anthropic.com · · Covered in 13 articles from 10 sources

The 98% Problem: A Survey of Harness Engineering for AI Agents

labs.beconfident.app··Hacker News

The Anthropic leader who built Claude Code says he ditched prompting — now he just writes loops.

thenewstack.io·

keyuchen21/agentic-engineering-handbook: The definitive OpenAI, Claude, MCP, Harness, Evals, and Production Agent Systems learning roadmap.

github.com··Hacker News

Long-Running Agents

modulecollective/moe: Ministry of Everything (MoE) is an anti-social agent harness for a single operator and a gaggle of bots.

github.com··Hacker News

I built a (very small) long running agent

roundup.getdbt.com··Hacker News

AI Agent Failure Modes Beyond Hallucination

Learn Harness Engineering에 오신 것을 환영합니다

walkinglabs.github.io·

Welcome to Learn Harness Engineering

walkinglabs.github.io··Hacker News

All Agent Harnesses: The Live Comparison

In other languages

どこからでもAIで仕事をできるようにしたけど、疲れてやめた話

Чем умнее модель, тем меньше ей нужно: четыре дисциплины production‑агента

Harnessing des agents : 9 pratiques venant du logiciel

blog.octo.com·