[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-74186":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":10,"archived":22,"fork":22,"defaultBranch":23,"hasWiki":24,"hasPages":24,"topics":25,"createdAt":10,"pushedAt":10,"updatedAt":26,"readmeContent":27,"aiSummary":28,"trendingCount":16,"starSnapshotCount":16,"syncStatus":29,"lastSyncTime":30,"discoverSource":31},74186,"harness-books","wquguru\u002Fharness-books","wquguru","📚 Two books on harness engineering — the design philosophies behind Claude Code & Codex: constraints, query loops, context governance, multi-agent verification. harness-books.agentway.dev","http:\u002F\u002Fharness-books.agentway.dev\u002F",null,"Python",2445,286,10,1,0,39,127,505,117,29.37,false,"main",true,[],"2026-06-12 02:03:23","# Harness Books\n\n[中文 README](.\u002FREADME.zh-CN.md)\n\n[![Read Online](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FRead%20Online-Harness%20Books-16a34a?style=flat-square&logo=googlechrome&logoColor=white)](https:\u002F\u002Fharness-books.agentway.dev\u002Fen\u002F)\n[![About AgentWay](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FAbout-AgentWay-22c55e?style=flat-square&logo=bookstack&logoColor=white)](https:\u002F\u002Fagentway.dev)\n\nTwo books on harness engineering. They pursue the same engineering question: once a code-writing model is placed inside terminals, repositories, permission systems, and team workflows, what keeps the overall system bounded, continuous, and accountable for consequences?\n\n\u003Ctable>\n  \u003Ctr>\n    \u003Ctd align=\"center\" valign=\"top\" width=\"50%\">\n      \u003Ca href=\"https:\u002F\u002Fharness-books.agentway.dev\u002Fen\u002Fbook1-claude-code\u002F\">\n        \u003Cimg src=\".\u002Fbook1-claude-code\u002Fassets\u002Fcover-wxb-en.svg\" alt=\"Harness Engineering: A Design Guide to Claude Code\" width=\"280\">\n      \u003C\u002Fa>\n      \u003Cbr>\n      \u003Cstrong>Harness Engineering: A Design Guide to Claude Code\u003C\u002Fstrong>\n      \u003Cbr>\n      \u003Ca href=\"https:\u002F\u002Fharness-books.agentway.dev\u002Fen\u002Fbook1-claude-code\u002F\">Read online\u003C\u002Fa> ·\n      \u003Ca href=\"https:\u002F\u002Fharness-books.agentway.dev\u002Fen\u002Fbook1-claude-code\u002Fexported\u002Fbook1-claude-code-en.pdf\">Download PDF\u003C\u002Fa>\n    \u003C\u002Ftd>\n    \u003Ctd align=\"center\" valign=\"top\" width=\"50%\">\n      \u003Ca href=\"https:\u002F\u002Fharness-books.agentway.dev\u002Fen\u002Fbook2-comparing\u002F\">\n        \u003Cimg src=\".\u002Fbook2-comparing\u002Fassets\u002Fcover-wxb-en.svg\" alt=\"The Harness Design Philosophies of Claude Code and Codex\" width=\"280\">\n      \u003C\u002Fa>\n      \u003Cbr>\n      \u003Cstrong>The Harness Design Philosophies of Claude Code and Codex\u003C\u002Fstrong>\n      \u003Cbr>\n      \u003Ca href=\"https:\u002F\u002Fharness-books.agentway.dev\u002Fen\u002Fbook2-comparing\u002F\">Read online\u003C\u002Fa> ·\n      \u003Ca href=\"https:\u002F\u002Fharness-books.agentway.dev\u002Fen\u002Fbook2-comparing\u002Fexported\u002Fbook2-comparing-en.pdf\">Download PDF\u003C\u002Fa>\n    \u003C\u002Ftd>\n  \u003C\u002Ftr>\n\u003C\u002Ftable>\n\nThese books are not meant to walk line by line through the source code. They focus on how a harness organizes constraints and execution, and how an inherently unstable model can be folded into a sustainable engineering order. Prompt layering, query loops, permission decisions, context governance, failure recovery, multi-agent verification, local rules, and team institutions together form the organ system of a harness. The real danger is not that a model occasionally says something wrong, but that the system has no structure for handling consequences.\n\n## Core Claims\n\n- Harness engineering is about how constraint structures organize execution.\n- Once a code-writing model enters a real engineering environment, the main problem is no longer answer quality but behavioral consequences.\n- Prompts, tools, permissions, state, recovery, verification, and institutions are not accessories around the system. They are organs in the same control structure.\n- When comparing agent systems, the key question is not the feature checklist but where order is actually placed.\n- If a team cannot turn individual experience into reusable rules, it will struggle to turn an agent into a stable system.\n\n## What The Two Books Focus On\n\n### Book 1: A Design Guide to Claude Code\n\nThe first book uses Claude Code as the observation target and concentrates on runtime structure. Its concern is why a system eventually has to grow components such as a control plane, query loop, tool permissions, context governance, recovery paths, multi-agent verification, and team rules.\n\nStart with Book 1 if these are the questions you care about:\n\n- Why harness engineering is not just prompt engineering at a larger scale\n- Why prompts are fundamentally part of the control plane rather than a chat box\n- Why model mistakes should be treated as a runtime norm rather than an exceptional event\n- Why multi-agent work and verification should not be blended into one vague mechanism\n- How a team can solidify personal experience into reusable engineering institutions\n\n### Book 2: Comparing Claude Code and Codex\n\nThe second book places Claude Code and Codex side by side and asks where each harness places order. One path starts from runtime discipline; the other starts from a more structured control layer. Both systems can work, but they distribute authority differently.\n\nStart with Book 2 if you care more about system choice, architectural judgment, or what to learn when building your own harness:\n\n- What the biggest control-plane divergence is between Claude Code and Codex\n- How to align the roles of query loops, threads, rollouts, and state\n- What governance roles are played by permissions, sandboxes, and policy languages\n- How skills, hooks, and local rules encode organizational habits into the system\n- If you want to build your own harness, who to learn from first and which layer to study first\n\n## Suggested Reading Paths\n\n- Want the full frame first: read Book 1, then Book 2.\n- Already familiar with coding-agent tools and want the architectural split directly: start with Book 2.\n- Only want the conclusions: read Book 1 Chapter 9 plus Book 2 Chapter 7.\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Full Table of Contents\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n### Book 1 — Harness Engineering: A Design Guide to Claude Code\n\n- [Introduction](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Findex.md)\n- [Preface: Harness, Terminals, and Engineering Constraints](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fpreface.md)\n- [Chapter 1 Why Harness Engineering Matters](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fchapter-01-why-harness-engineering.md)\n- [Chapter 2 Prompt Is Not Personality, Prompt Is the Control Plane](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fchapter-02-prompt-is-control-plane.md)\n- [Chapter 3 Query Loop: The Heartbeat of an Agent System](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fchapter-03-query-loop-heartbeat.md)\n- [Chapter 4 Tools, Permissions, and Interrupts: Why Agents Cannot Touch the World Directly](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fchapter-04-tools-permissions-interrupts.md)\n- [Chapter 5 Context Governance: Memory, CLAUDE.md, and Compact as a Budgeting Regime](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fchapter-05-context-memory-compact.md)\n- [Chapter 6 Errors and Recovery: An Agent System That Keeps Working After Failure](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fchapter-06-errors-and-recovery.md)\n- [Chapter 7 Multi-Agent Work and Verification: Managing Instability Through Division of Labor](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fchapter-07-multi-agent-and-verification.md)\n- [Chapter 8 Team Adoption: Turning a Smart Tool into a Reusable Institution](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fchapter-08-team-landing-practices.md)\n- [Chapter 9 Ten Principles of Harness Engineering](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fchapter-09-ten-principles.md)\n- [Appendix A Checklists: Turning Principles into Executable Constraints](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fappendix-a-checklists.md)\n- [Appendix B Diagrams: Drawing the Runtime Skeleton](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fappendix-b-diagram-notes.md)\n- [Appendix C Source Map: Which Files Ground Each Chapter](.\u002Fbook1-claude-code\u002Flocales\u002Fen\u002Fappendix-c-source-map.md)\n\n### Book 2 — The Harness Design Philosophies of Claude Code and Codex\n\n- [Introduction](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Findex.md)\n- [Reading Map: How to Understand Book 1 and This Comparative Book Together](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fchapter-00-reading-map.md)\n- [Preface: Two Harnesses, Not Accessories on the Same Horse](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fpreface.md)\n- [Chapter 1: Why We Compare Claude Code and Codex](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fchapter-01-why-this-comparison.md)\n- [Chapter 2: Two Control Planes: Prompt Assembly and Instruction Fragments](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fchapter-02-two-control-planes.md)\n- [Chapter 3: Where the Heartbeat Lives: Query Loop Compared with Thread, Rollout, and State](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fchapter-03-loop-thread-and-rollout.md)\n- [Chapter 4: Tools, Sandboxes, and Policy Languages: Who Stops the Model from Moving Too Fast](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fchapter-04-tools-sandbox-and-exec-policy.md)\n- [Chapter 5: Skills, Hooks, and Local Rules: How a System Learns Local Discipline](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fchapter-05-skills-hooks-and-local-governance.md)\n- [Chapter 6: Delegation, Verification, and Persistent State: Who Prevents a System from Grading Itself](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fchapter-06-delegation-verification-and-state.md)\n- [Chapter 7: Convergence Through Different Roads, or Separate Branches](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fchapter-07-convergence-and-divergence.md)\n- [Chapter 8: If You Are Building Your Own Harness, What to Study First](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fchapter-08-how-to-choose-or-build.md)\n- [Appendix A: Source Map That Underpins the Comparison](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fappendix-a-source-map.md)\n- [Appendix B: Checklist to Determine Where Your Harness Sits](.\u002Fbook2-comparing\u002Flocales\u002Fen\u002Fappendix-b-checklists.md)\n\n\u003C\u002Fdetails>\n\n## Want To Keep Practicing? Try AgentWay\n\n\u003Ctable>\n\u003Ctr>\n\u003Ctd width=\"180\" align=\"center\" valign=\"middle\">\n  \u003Ca href=\"https:\u002F\u002Fagentway.dev\u002F\">\n    \u003Cimg src=\"assets\u002Fagentway-logo.svg\" alt=\"AgentWay\" width=\"150\">\n  \u003C\u002Fa>\n\u003C\u002Ftd>\n\u003Ctd valign=\"middle\">\n  \u003Cb>\u003Ca href=\"https:\u002F\u002Fagentway.dev\u002F\">AgentWay\u003C\u002Fa>\u003C\u002Fb> is a related but separate practice platform. Harness Books explains control structures, engineering judgments, and architectural divergences. AgentWay is where these ideas continue into training paths, drills, project exercises, and agent PoCs.\n\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftable>\n\n## Local Build\n\nBuild the two locale-aware Honkit sites and then assemble the unified Pages site:\n\n```bash\npython3 tools\u002Fbook-kit\u002Fbuild_honkit.py book1-claude-code\npython3 tools\u002Fbook-kit\u002Fbuild_honkit.py book1-claude-code --locale en\npython3 tools\u002Fbook-kit\u002Fbuild_honkit.py book2-comparing\npython3 tools\u002Fbook-kit\u002Fbuild_honkit.py book2-comparing --locale en\npython3 tools\u002Fbook-kit\u002Fbuild_pages_site.py\n```\n\nFinal output is written to `dist\u002F`.\n\n---\n\n\u003Csub>Keywords: Harness Engineering, Claude Code guide, Claude Code vs Codex, AI coding agent, control plane, query loop, agent recovery, agent verification, local governance, approval policy\u003C\u002Fsub>\n","该项目提供了两本关于约束工程的书籍，旨在探讨当代码编写模型被置于终端、仓库、权限系统和团队工作流中时，如何保持整个系统的边界性、连续性和对后果的责任。核心功能和技术特点包括通过约束结构组织执行、处理行为后果而非仅仅关注答案质量、以及将提示、工具、权限、状态恢复等视为同一控制结构中的器官。这些内容特别适合需要理解和实践如何在实际工程环境中管理AI模型行为及其影响的软件开发团队或个人工程师阅读。",2,"2026-06-11 03:49:25","high_star"]