惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - 叶小钗
O
OpenAI News
V
V2EX
大猫的无限游戏
大猫的无限游戏
博客园 - 聂微东
S
Schneier on Security
C
CXSECURITY Database RSS Feed - CXSecurity.com
小众软件
小众软件
L
LINUX DO - 热门话题
C
Cybersecurity and Infrastructure Security Agency CISA
博客园 - Franky
Security Latest
Security Latest
S
SegmentFault 最新的问题
Project Zero
Project Zero
Spread Privacy
Spread Privacy
K
Kaspersky official blog
J
Java Code Geeks
V
Vulnerabilities – Threatpost
C
Cisco Blogs
C
CERT Recently Published Vulnerability Notes
月光博客
月光博客
T
The Exploit Database - CXSecurity.com
L
Lohrmann on Cybersecurity
人人都是产品经理
人人都是产品经理
博客园 - 三生石上(FineUI控件)
Scott Helme
Scott Helme
WordPress大学
WordPress大学
量子位
T
Threat Research - Cisco Blogs
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
宝玉的分享
宝玉的分享
Hugging Face - Blog
Hugging Face - Blog
AWS News Blog
AWS News Blog
Help Net Security
Help Net Security
Application and Cybersecurity Blog
Application and Cybersecurity Blog
Simon Willison's Weblog
Simon Willison's Weblog
S
Secure Thoughts
博客园 - 【当耐特】
cs.CV updates on arXiv.org
cs.CV updates on arXiv.org
V
Visual Studio Blog
Last Week in AI
Last Week in AI
T
Tailwind CSS Blog
腾讯CDC
Cyberwarzone
Cyberwarzone
IT之家
IT之家
GbyAI
GbyAI
Exploit-DB.com RSS Feed
Exploit-DB.com RSS Feed
云风的 BLOG
云风的 BLOG
T
Troy Hunt's Blog
D
Docker

博客园 - lightsong

Micro Frontends extending the microservice idea to frontend development What is Nx? Smart Monorepo Build System & CI Build and Query Knowledge Graphs with LLMs WizMap AI Code Review Claude Code Awesome Claude Agents - 26-agent AI Development Team (open source) WezTerm Herdr Zellij playwright-cli branch vs worktree Ralph Loop Browserbase Skills The Complete Beginner's Guide to GSD (Get Shit Done) Framework for Claude Code A Claude Code Skills Stack: How to Combine Superpowers, gstack, and GSD Without the Chaos Casdoor The 8 Levels of Agentic Engineering assistant-ui vs copilotkit vs tambo Apache Casbin Agentic AI-Powered HR Automation Instant CV Intelligence for Modern Hiring Teams Python + LangGraph + FastAPI langfuse contextvars Agent Skills Temporal + LangGraph: A Two-Layer Architecture for Multi-Agent Coordination Building Generative AI Services with FastAPI DeepAgent vs LangGraph Plan-and-Execute AI-Native Software Delivery 动态表单EAV模型(Entity-Attribute-Value模型)和JSON字段 ClickHouse® is a real-time analytics database management system GRAPH RAG 数仓建模——维度表详细讲解 涌现能力 The 5W2H Problem-Solving Method 前端如何直接上传文件夹 docker状态混乱修复(如果出现容器stop不了问题,报Permission denied) Ducky - BPMN 工作流管理系统
Harness engineering for coding agent users
lightsong · 2026-04-22 · via 博客园 - lightsong

Harness engineering for coding agent users

https://openai.com/zh-Hans-CN/index/harness-engineering/

https://www.anthropic.com/engineering/harness-design-long-running-apps

https://martinfowler.com/articles/harness-engineering.html

 Agent = Model + Harness

image

 A well-built outer harness serves two goals: it increases the probability that the agent gets it right in the first place, and it provides a feedback loop that self-corrects as many issues as possible before they even reach human eyes. Ultimately it should reduce the review toil and increase the system quality, all with the added benefit of fewer wasted tokens along the way.

image

image

image

image

image

image

image

image

image

出处:http://www.cnblogs.com/lightsong/ 本文版权归作者和博客园共有,欢迎转载,但未经作者同意必须保留此段声明,且在文章页面明显位置给出原文连接。